A model of liver injury caused by echinococcosis and a lesion tissue positioning method

Abstract

The application discloses a liver damage model caused by echinococcosis and a lesion tissue positioning method, which comprises an image acquisition module, a processor, a memory and a display which are connected with each other. On the basis of a three-dimensional liver ultrasonic image, after image gray processing, convolution operation is carried out, and the three-dimensional image is divided into two-dimensional images which are easy to process. Through feature extraction and data type merging, division of different liver tissue regions is realized, and whether liver echinococcosis is alive or not can be accurately judged, meanwhile, false points are eliminated, lesion tissue judgment efficiency is higher, precise positioning of lesion tissue can be realized, judgment can be realized without the help of professional knowledge of medical staff, treatment efficiency is high, and surgical risk is low.

Description

Technical Field

[0001] This invention belongs to the field of liver disease detection technology, specifically relating to a liver injury model caused by echinococcosis and a method for locating lesions. Background Technology

[0002] Hepatic echinococcosis is a common parasitic disease in grassland pastoral areas, affecting both humans and animals. The vast majority of cases are caused by the ingestion of canine tapeworm eggs, which hatch into oncocytozoa in the stomach and duodenum, then travel via the portal vein to the liver and develop into echinococcosis cysts. A minority are caused by echinococcosis, resulting from the larval stages of *Echinococcus vesicatus*. Hepatic echinococcosis most commonly affects the right lobe, with fewer cases in the left lobe or both lobes. Clinically, it is divided into two types: ① Unilocular echinococcosis (echinococcosis cysts or cystic echinococcosis): This presents as a slow-growing, encapsulated cyst. Daughter cysts grow from the inner wall (germinal layer) of the cyst into the lumen, and scolex grows from the inner wall of the daughter cysts. ② Alveolar echinococcosis: This forms small, hard, grayish-white vesicles in the liver, containing no cystic fluid but a curd-like substance. Both types of echinococcosis cause symptoms such as upper abdominal discomfort, loss of appetite, nausea, and vomiting.

[0003] In recent years, the diagnosis of hepatic echinococcosis based on ultrasound image classification and analysis, and the surgical resection of lesions based on etiological judgment and lesion tissue images, have been widely applied in clinical practice, achieving certain positive results. However, with current techniques, medical personnel still rely on preliminary judgments from ultrasound images when resecting lesions. It is not easy to determine the boundary between lesions and normal hepatocytes, the location of blood vessels, or whether the lesion has been completely removed. Therefore, current methods cannot accurately guide surgical procedures, resulting in significant surgical risks and frequent cases of recurrence due to incomplete resection.

[0004] Based on this, the present invention aims to establish a liver injury model and lesion localization method caused by echinococcosis on the basis of existing technology, so as to better assist medical personnel in accurately determining the location and cause of lesions, improve the efficiency of clinical diagnosis and treatment of echinococcosis, and reduce surgical risks. Summary of the Invention

[0005] To address the technical challenges identified in the background section, this invention proposes a liver injury model caused by echinococcosis and a method for locating lesions. The specific technical solutions for the model and its separation device are as follows:

[0006] A liver injury model caused by echinococcosis includes an image acquisition module, a processor, a memory, and a display. The processor is electrically connected to the image acquisition module, the memory, and the display. The image acquisition module is used to acquire three-dimensional images of the liver. The processor is used to perform grayscale, slicing, convolution, and merging processing on the three-dimensional liver images. The processor further includes an image segmentation unit, a feature extraction module, a data processing module, a sub-unit module, and an information integration module connected in sequence. The input end of the image segmentation unit is connected to the output end of the image acquisition module, and the output end of the information integration module is connected to the input end of the display. The display is used to display only the three-dimensional segmented image processed by the processor.

[0007] The image acquisition module acquires three-dimensional images of the liver at regular intervals using a three-dimensional ultrasound imager. The three-dimensional images of the liver are used to establish the location region of the liver and distinguish the left and right lobes of the liver.

[0008] The image segmentation unit processes the grayscale of the three-dimensional image of the liver, and then segments the three-dimensional image of the liver through a convolution calculation module and a convolutional network unit to obtain multiple two-dimensional slice images of the liver.

[0009] The feature extraction module divides the two-dimensional slice image into multiple pixel units, extracts corner points, edges, and grayscale features, and obtains multiple sets of feature tensor data, such as data1, data2, ..., dataN. At the same time, the position coordinate data of the corresponding pixel units, po1(x1,y1,z1), po2(x2,y2,z2), ..., poN(xN,yN,zN), are incorporated into the feature tensor data to obtain a first format dataset Dt[Dt1,Dt1,...,DtN]. The first format dataset is then uploaded to the data processing module.

[0010] The data processing module trains convolutional layers on the elements in the first format dataset. During the training process, each element in the first format dataset is compared. That is, under the condition that the feature tensor data is the same, it is determined whether the position coordinates are adjacent. If they are adjacent, they are merged into the same second format dataset. There are a total of four second format datasets: hepatobiliary duct format dataset H1, blood vessel format dataset H2, gallbladder format dataset H3, and lesion tissue format dataset H4.

[0011] The training expression for the convolutional layers of the above convolutional network is: Dt m+1 =f(Dt) m ,t,K), where K represents the number of the two-dimensional slice image, m is the pixel size, and t is the sampling time.

[0012] The sub-unit module comprises four sub-units: a hepatobiliary duct sub-unit, a blood vessel sub-unit, a gallbladder sub-unit, and a diseased tissue sub-unit. This sub-unit module is used to model various tissues of the liver. Each sub-unit utilizes the aforementioned second-format dataset to model different structural tissues of the liver. The second-format dataset is transmitted to each sub-unit, where the physical connections between data points form surface regions. All surface regions are then represented in grayscale to reconstruct the different sub-unit modules.

[0013] The hepatobiliary duct sub-module contains the three-dimensional structure of the hepatobiliary ducts; the blood vessel sub-module is the three-dimensional structure of blood vessels, thus showing the distribution pattern of blood vessels inside the liver; the gallbladder sub-module shows the location and shape of the gallbladder; and the lesion tissue sub-module shows the distribution pattern and distribution area of ​​echinococcosis lesion tissue inside the liver.

[0014] The sub-unit module also includes a liver proton sub-unit, which uses a 3dVR algorithm to process the transparency of the original three-dimensional liver image to obtain a transparent three-dimensional liver proton image.

[0015] The information integration module, based on the completion of each sub-unit module, fuses the data of the sub-unit modules to form the final region segmentation result, namely the liver region image, and only the region affected by echinococcosis is represented by RGB true color.

[0016] Based on the above, the final liver segmentation image fused by the information integration module will be displayed on a monitor. Medical staff will then formulate a surgical strategy and perform the surgery based on the three-dimensional segmented liver image displayed on the monitor.

[0017] Based on the echinococcosis-induced liver injury model, this invention further proposes a lesion tissue localization method based on the echinococcosis-induced liver injury model, the localization method comprising the following steps:

[0018] Step S1: The image acquisition module acquires three-dimensional ultrasound images of the patient's liver at regular intervals and transmits the three-dimensional ultrasound images to the image segmentation unit;

[0019] Step S2: The image segmentation unit processes the grayscale of the three-dimensional image of the liver, and segments the three-dimensional image of the liver through the convolution calculation module and the convolution network unit to obtain multiple two-dimensional slice images of the liver.

[0020] Step S3: Divide the two-dimensional slice image into multiple pixel units, extract corner points, edges, and grayscale features to obtain multiple sets of feature tensor data, such as data1, data2, ..., dataN. At the same time, incorporate the position coordinate data of the corresponding pixel units, po1(x1,y1,z1), po2(x2,y2,z2), ..., poN(xN,yN,zN), into the feature tensor data to obtain the first format dataset Dt[Dt1,Dt1,......,DtN];

[0021] Step S4: Identify artifact points and construct an artifact sparse matrix A using the artifact points in the 2D slice image. ij Perform sequential storage and Boolean operations with the first format dataset Dt (i.e., Dt&A). ij The sparse matrix A of artifacts is replaced with zero elements. ij The non-zero elements in the dataset are used to update the first format dataset Dt to obtain Dt', and then the artifacts are removed. The updated first format dataset Dt' is then uploaded to the data processing module.

[0022] Step S5: Train convolutional layers for the elements in the updated first format dataset, compare each element in the first format dataset, and determine whether the position coordinates are adjacent under the condition that the feature tensor data are the same. If they are adjacent, merge them into the same second format dataset.

[0023] The second format dataset includes: hepatobiliary duct format dataset H1, vascular format dataset H2, gallbladder format dataset H3, and lesion tissue format dataset H4;

[0024] Step S6: The sub-unit module uses the second format datasets H1, H2, H3, and H4 to perform three-dimensional data modeling, so that adjacent data points are physically connected to form surface regions. All surface regions are filled with grayscale to restore different sub-unit modules: hepatobiliary duct sub-unit, blood vessel sub-unit, gallbladder sub-unit, and diseased tissue sub-unit.

[0025] Step S7: Based on step S6, the sub-unit module includes a liver proteome sub-unit, which uses a 3dVR algorithm to make the liver proteome image transparent.

[0026] Step S8: The information integration module merges the data from each sub-unit module to form a liver region image. In the liver region image, only the areas affected by echinococcosis are represented using RGB true color and displayed on the monitor.

[0027] In summary, the liver injury model and lesion localization method caused by echinococcosis of the present invention differ from the prior art in that:

[0028] 1) Based on three-dimensional liver ultrasound images, this invention processes the images in grayscale and performs convolution operations to segment the three-dimensional images into two-dimensional images that are easy to process. Through feature extraction and data type merging, it realizes the division of different tissue regions of the liver, making the identification of lesions more efficient and enabling identification without the need for medical expertise.

[0029] 2) This invention deconstructs three-dimensional liver ultrasound images into different sub-unit modules, and constructs three-dimensional maps of each tissue in the sub-unit modules. This enables medical personnel to identify the distribution and specific location of diseased tissues, achieve precise localization of diseased tissues, improve the efficiency of clinical diagnosis and treatment of echinococcosis, and reduce surgical risks.

[0030] 3) This invention compares grayscale image data of lesion tissue at different times, which can accurately determine whether the liver echinococcosis is alive, and then take corresponding measures;

[0031] 4) This invention can effectively eliminate artifacts in three-dimensional images, ensuring the accuracy of three-dimensional image recognition of the liver. Attached Figure Description

[0032] Figure 1 This is a diagram illustrating the modeling process of a liver injury model caused by echinococcosis according to the present invention.

[0033] Figure 2 for Figure 1 (d) Detailed diagram of sub-unit module;

[0034] In the diagram, 1-hepatic bile duct; 2-blood vessel; 3-gallbladder; 4-hydatid disease lesion tissue; 5-subunit module. Detailed Implementation

[0035] The technical solution of the present invention will be further described and explained below with reference to the accompanying drawings and specific embodiments, in order to help those skilled in the art to understand the present invention more deeply and completely.

[0036] Example 1

[0037] Please investigate. Figure 1 As shown, the liver injury model caused by echinococcosis of the present invention includes: an image acquisition module, a processor, a memory, and a display. The processor is electrically connected to the image acquisition module, the memory, and the display. The processor further includes an image segmentation unit, a feature extraction module, a data processing module, a sub-unit module, and an information integration module connected in sequence. The input end of the image segmentation unit is connected to the output end of the image acquisition module, and the output end of the information integration module is connected to the input end of the display.

[0038] The image acquisition module acquires three-dimensional images of the liver at regular intervals using a three-dimensional ultrasound imager (e.g., Figure 1 (as shown in (a)) The three-dimensional image of the liver is used to establish the location region of the liver and distinguish the left and right lobes of the liver;

[0039] The image segmentation unit processes the grayscale of the three-dimensional image of the liver, and then segments the three-dimensional image of the liver through the convolution calculation module and the convolution network unit to obtain multiple two-dimensional slice images of the liver, as shown in Figure (b).

[0040] The feature extraction module divides the two-dimensional slice image into multiple pixel units (e.g., converting the two-dimensional slice image into a grayscale image unit of 1680*2100 pixels), extracts corner points, edges, and grayscale features, and obtains multiple sets of feature tensor data, such as data1, data2, ..., dataN. At the same time, the position coordinate data of the corresponding pixel units, po1(x1,y1,z1), po2(x2,y2,z2), ..., poN(xN,yN,zN), are incorporated into the feature tensor data to obtain the first format dataset Dt[Dt1,Dt1,...,DtN], which is then uploaded to the data processing module.

[0041] The data processing module trains convolutional layers on the elements in the first format dataset. During the training process, each element in the first format dataset is compared. That is, under the condition that the feature tensor data is the same, it is determined whether the position coordinates are adjacent. If they are adjacent, they are merged into the same second format dataset. There are a total of four second format datasets: hepatobiliary duct format dataset H1, blood vessel format dataset H2, gallbladder format dataset H3, and lesion tissue format dataset H4.

[0042] The training expression for the convolutional layers of the above convolutional network is: Dt m+1 =f(Dt) m ,t,K), where K represents the number of the two-dimensional slice image, m is the pixel size, and t is the sampling time.

[0043] The sub-unit module contains four sub-units: the hepatobiliary duct sub-unit, the blood vessel sub-unit, the gallbladder sub-unit, and the diseased tissue sub-unit. These are used to model various tissues within the liver. Each sub-unit utilizes the aforementioned second-format dataset to model different structural tissues of the liver. The second-format dataset is transmitted to the sub-units, where physical connections between data points form surface regions. Grayscale representation of all surface regions is then used to reconstruct different 3D data sub-models, i.e., the sub-unit module. Figure 1 As shown in (d).

[0044] Further investigation is requested. Figure 2 As shown, the hepatobiliary duct subunit module contains the three-dimensional structure of hepatobiliary duct 1. Figure 2(Top left image); the vascular subunit is the three-dimensional structure of vessel 2 (including hepatic veins, hepatic arteries, and portal veins). Figure 2 The upper right image shows the distribution of blood vessels within the liver; the gallbladder sub-unit module shows the location and morphology of gallbladder 3. Figure 2 (Lower left figure); while the lesion tissue subunit shows the distribution pattern and distribution area of ​​echinococcosis lesion tissue 4 within the liver ( Figure 2 (bottom right image).

[0045] In addition, the sub-unit module also includes a liver proton sub-unit. The liver proton sub-unit uses the 3dVR algorithm (i.e., the three-dimensional voxel algorithm) to process the transparency of the original three-dimensional liver image to obtain a transparent three-dimensional liver image. The opacity coefficient γ is set between 0.2 and 0.35.

[0046] The information integration module, based on the completion of each sub-module, fuses the data from these sub-modules to form the final region segmentation result, i.e., the liver region image, such as... Figure 1 As shown in (d), only the areas affected by echinococcosis are represented in RGB true color.

[0047] The above data fusion is based on the fact that all sub-unit modules have the same volume / size and the same coordinate origin. During data fusion, the 3D grayscale images from multiple sub-unit modules with the same coordinate origin are superimposed.

[0048] In addition, since the image acquisition module periodically acquires three-dimensional images of the liver, the comparator in the processor can compare the image of the lesion tissue subunit formed at time T2 with the image of the lesion tissue subunit formed at time T1. If the RGB true color of a certain area changes, it indicates that the liver hydatid is in a growth state in that area.

[0049] Based on this, the final liver segment image fused by the information integration module will be displayed on a monitor. Medical staff will then formulate a surgical strategy and perform the surgery based on the three-dimensional segmented liver image displayed on the monitor.

[0050] Example 2

[0051] A method for lesion tissue localization based on a liver injury model caused by echinococcosis, the method comprising the following steps:

[0052] Step S1: The image acquisition module acquires three-dimensional ultrasound images of the patient's liver at regular intervals and transmits the three-dimensional ultrasound images to the image segmentation unit;

[0053] Step S2: The image segmentation unit processes the grayscale of the three-dimensional image of the liver, and segments the three-dimensional image of the liver through the convolution calculation module and the convolution network unit to obtain multiple two-dimensional slice images of the liver.

[0054] Step S3: Divide the two-dimensional slice image into multiple pixel units, extract corner points, edges, and grayscale features to obtain multiple sets of feature tensor data, such as data1, data2, ..., dataN. At the same time, incorporate the position coordinate data of the corresponding pixel units, po1(x1,y1,z1), po2(x2,y2,z2), ..., poN(xN,yN,zN), into the feature tensor data to obtain the first format dataset Dt[Dt1,Dt1,......,DtN];

[0055] Step S4: Identify artifact points and construct an artifact sparse matrix A using the artifact points in the 2D slice image. ij Perform sequential storage and Boolean operations with the first format dataset Dt (i.e., Dt&A). ij The sparse matrix A of artifacts is replaced with zero elements. ij The non-zero elements in the dataset are used to update the first format dataset Dt to obtain Dt', and then the artifacts are removed. The updated first format dataset Dt' is then uploaded to the data processing module.

[0056] Step S5: Train convolutional layers for the elements in the updated first format dataset, compare each element in the first format dataset, and determine whether the position coordinates are adjacent under the condition that the feature tensor data are the same. If they are adjacent, merge them into the same second format dataset.

[0057] The second format dataset includes: hepatobiliary duct format dataset H1, vascular format dataset H2, gallbladder format dataset H3, and lesion tissue format dataset H4;

[0058] Step S6: The sub-unit module uses the second format datasets H1, H2, H3, and H4 to perform three-dimensional data modeling, so that adjacent data points are physically connected to form surface regions. All surface regions are filled with grayscale to restore different sub-unit modules: hepatobiliary duct sub-unit, blood vessel sub-unit, gallbladder sub-unit, and diseased tissue sub-unit.

[0059] Step S7: Based on step S6, the sub-unit module includes a liver proteome sub-unit, which uses a 3dVR algorithm to make the liver proteome image transparent.

[0060] Step S8: The information integration module merges the data from each sub-unit module to form a liver region image. In the liver region image, only the areas affected by echinococcosis are represented using RGB true color and displayed on the monitor.

[0061] Example 3

[0062] Based on step S5 in Example 2, the following steps S51 to S52 are also included:

[0063] Step S51: Upload the lesion tissue format dataset H4 in the second format dataset to the comparator every time interval ΔT, compare the lesion tissue sub-unit image at the current time with the lesion tissue sub-unit image at the previous time, and determine whether there is a change in the RGB true color of the region.

[0064] Step S52: If the RGB true color of a certain region changes twice within 3 consecutive ΔT, it is determined that the hydatid cyst in that region is in a growing state; if the RGB true color of a certain region does not change within 3 consecutive ΔT, it is determined that the hydatid cyst in that region is in a dead state.

[0065] The ΔT is taken as 10 min to 15 min (including 10 min and 15 min).

[0066] In summary, the present invention provides a lesion tissue localization method based on a liver injury model caused by echinococcosis. Using three-dimensional ultrasound images and employing multi-layer convolutional operations for grayscale processing and classification, the entire liver is segmented and separated. This method can efficiently control the accuracy and speed of the segmentation process, resulting in highly accurate segmentation results. It is more conducive to guiding medical personnel in precise operations, improving diagnostic efficiency, and is not limited by individual differences in liver morphology, thus having a broader application prospect in the field of liver disease diagnosis.

[0067] Obviously, those skilled in the art can make various modifications or simple transformations to the technical solutions of this invention without departing from the spirit and scope of this invention. These modifications or simple transformations, which are the result of no inventive effort, should be considered as included within the scope defined by the claims of this invention.

Claims

1. A method for locating a lesion of liver damage caused by echinococcosis, characterized by, The positioning method comprises the following steps: Step S1: acquiring a three-dimensional ultrasound image of a patient's liver every interval of time by an image acquisition module, and delivering the three-dimensional ultrasound image to an image segmentation unit; Step S2: the image segmentation unit processes the liver three-dimensional image in grayscale, segments the liver three-dimensional image through a convolution calculation module and a convolution network unit, and obtains a plurality of liver two-dimensional slice images; Step S3: the two-dimensional slice images are divided to form a plurality of pixel units, the corner points, edges and grayscale features are extracted, a plurality of groups of feature tensor data data1, data2,..., dataN are obtained, and the position coordinate data po1(x1, y1, z1), po2(x2, y2, z2),..., poN(xN, yN, zN) of the corresponding pixel units are fused into the feature tensor data to obtain a first format data set Dt[Dt1, Dt2,..., DtN]; Step S4: Identify artifact points and construct an artifact sparse matrix A using the artifact points in the 2D slice image. ij The data is stored sequentially, and Boolean operations are performed with the first-format dataset Dt. The artifact sparse matrix A is replaced with zero elements. ij The non-zero elements in the dataset are used to update the first format dataset Dt to obtain Dt', and then the artifacts are removed. The updated first format dataset Dt' is then uploaded to the data processing module. Step S5: the elements in the updated first format data set are subjected to convolution layer training, each element in the first format data set is compared, and under the condition that the feature tensor data is the same, it is judged whether the position coordinates are adjacent, and if adjacent, they are merged into the same second format data set; The second format data set comprises: a hepatic duct format data set H1, a blood vessel format data set H2, a gallbladder format data set H3 and a lesion tissue format data set H4; Step S6: the three-dimensional data modeling is performed by the sub-unit module using the second format data sets H1, H2, H3 and H4, the adjacent data points are physically connected to form a surface domain, the grayscale of all surface domains is filled, and different sub-unit modules are restored: a hepatic duct sub-unit, a blood vessel sub-unit, a gallbladder sub-unit and a lesion tissue sub-unit; Step S7: on the basis of step S6, the sub-unit module comprises a liver proton sub-unit, and the liver proton sub-unit performs transparent processing on the liver proton image by using a 3dVR algorithm; Step S8: the information integration module fuses the data of each sub-unit module to finally form a liver sub-region image, in which only the area affected by echinococcosis is represented by RGB true color, and the display is used for display.

2. A model of liver damage caused by echinococcosis, based on the method for locating lesions in tissue of liver damage caused by echinococcosis according to claim 1, characterized in that: The image acquisition module, the processor, the memory and the display, characterized in that: the processor is electrically connected with the image acquisition module, the memory and the display, the image acquisition module is used for acquiring a liver three-dimensional image, the processor is used for grayscale, slicing, convolution and merging processing of the liver three-dimensional image, the processor further comprises an image segmentation unit, a feature extraction module, a data processing module, a sub-unit module and an information integration module connected in sequence; the input end of the image segmentation unit is connected with the output end of the image acquisition module, the output end of the information integration module is connected with the input end of the display, and the display is used for displaying the three-dimensional segmented image processed by the processor.

3. The model for liver damage caused by echinococcosis according to claim 2, characterized in that: The image acquisition module acquires a liver three-dimensional image at a regular time through a three-dimensional ultrasound imaging instrument, and the liver three-dimensional image is used for establishing a liver position region and distinguishing a left liver lobe and a right liver lobe.

4. The model for liver damage caused by echinococcosis according to claim 3, characterized in that: The image segmentation unit segments the liver three-dimensional image to obtain a plurality of liver two-dimensional slice images by means of a convolution calculation module and a convolution network unit after gray processing of the liver three-dimensional image.

5. The model for liver damage caused by echinococcosis according to claim 4, characterized in that: The feature extraction module divides the two-dimensional slice image to form a plurality of pixel units, extracts corner points, edges and gray features to obtain a plurality of feature tensor data such as data1, data2,..., dataN, and simultaneously, integrates the position coordinate data po1(x1, y1, z1), po2(x2, y2, z2),..., poN(xN, yN, zN) of the corresponding pixel units into the feature tensor data to obtain first format data Dt[Dt1, Dt1,..., DtN], and uploads the first format data set to the data processing module.

6. The model for liver damage caused by echinococcosis according to claim 5, characterized in that: The data processing module performs convolution layer training on the elements in the first format data set, compares each element in the first format data set during the training process, that is, under the condition that the feature tensor data is the same, judges whether the position coordinates are adjacent, and if adjacent, merges into the same second format data set, a total of four second format data sets: a hepatic duct format data set H1, a blood vessel format data set H2, a gallbladder format data set H3 and a lesion tissue format data set H; Wherein, the convolution layer training expression of the convolution network is: Dt m+1 = f(Dt m , t, K), wherein K represents the number of two-dimensional slice images, m is the pixel size, and t is the sampling time.

7. The model for liver damage caused by echinococcosis according to claim 6, characterized in that: The subunit module includes four subunits: a hepatic duct subunit, a blood vessel subunit, a gallbladder subunit and a lesion tissue subunit, and the subunit module is used for modeling each tissue of the liver, and each subunit is used for modeling different structural tissues of the liver by means of the second format data set. The second format data set is transmitted to the subunit, and the data points are physically connected to form a surface domain, and all the surface domains are represented by gray scale to restore different subunit modules: a hepatic duct subunit module, a blood vessel subunit module, a gallbladder subunit module and a lesion tissue subunit module.

8. The model for liver damage caused by echinococcosis according to claim 7, characterized in that: The subunit module further includes a liver proton unit, which uses a 3dVR algorithm to process the transparency of the original liver three-dimensional image to obtain a transparent three-dimensional liver quality image.

9. The model for liver damage caused by echinococcosis according to claim 7, characterized in that: The information integration module integrates the data of the subunit modules to form a final region segmentation result, i.e., a liver subregion image, and only the region affected by hydatidosis is represented by RGB true color.

Images