Artificial intelligence algorithm-based laparoscopic liver resection operation registration method and system

Abstract

The invention provides a laparoscopic liver resection operation registration method and system based on an artificial intelligence algorithm, and belongs to the technical field of image processing. The method comprises the following steps: performing normalization processing on a preprocessed laparoscopic surgery image to obtain a normalized laparoscopic surgery image; performing contour extraction on the normalized laparoscopic surgery image by adopting an artificial intelligence algorithm to obtain an upper edge and a lower edge of the liver of the laparoscopic surgery image; constructing a preoperative 3D liver model, and marking the upper edge and the lower edge of the liver in the constructed preoperative 3D liver model; and performing registration calculation on the upper edge and the lower edge of the liver in the laparoscopic surgery image and the upper edge and the lower edge of the liver marked in the constructed preoperative 3D liver model to obtain a registration result. Compared with the prior art, the method effectively solves the problem that the accuracy of registration of the laparoscopic liver resection operation is not high.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of image processing, and particularly relates to a laparoscopic liver resection surgery registration method and system based on an artificial intelligence algorithm. BACKGROUND

[0002] With the continuous development of laparoscopic technology, monocular laparoscopes are increasingly widely applied in liver resection surgery. Due to the perspective limitation of laparoscopic images, there is a large deviation between the field of view in surgery and the three-dimensional structure of the liver, which further brings great difficulty to accurate positioning in surgery.

[0003] At present, most of the existing laparoscopic liver resection surgery registration methods depend on manual calibration or traditional visual navigation image registration algorithms. The above two methods not only require more prior knowledge and manual operation, but also lead to low accuracy of laparoscopic liver resection surgery registration, and are difficult to cope with complex liver morphology and perspective changes. SUMMARY

[0004] The purpose of the present application is to provide a laparoscopic liver resection surgery registration method and system based on an artificial intelligence algorithm, which solves the problem of low accuracy of laparoscopic liver resection surgery registration in the prior art.

[0005] In order to achieve the above-mentioned purpose, the present application adopts the following technical solutions: In a first aspect, the present application provides a laparoscopic liver resection surgery registration method based on an artificial intelligence algorithm, comprising the following steps: Obtaining a laparoscopic surgery image; Pretreating the obtained laparoscopic surgery image to obtain a pretreated laparoscopic surgery image, the pretreatment including noise removal and laparoscopic surgery image enhancement; Normalizing the pretreated laparoscopic surgery image to obtain a normalized laparoscopic surgery image; Extracting the contours of the normalized laparoscopic surgery image using an artificial intelligence algorithm to obtain the upper and lower edges of the liver in the laparoscopic surgery image; Constructing a preoperative 3D liver model and labeling the upper and lower edges of the liver in the constructed preoperative 3D liver model; Registering the upper and lower edges of the liver in the laparoscopic surgery image with the labeled upper and lower edges of the liver in the constructed preoperative 3D liver model to obtain a registration result.

[0006] A further improvement of the present application is that the artificial intelligence algorithm is an HRNet (High-Resolution Network) model.

[0007] A further improvement of the present invention is to use the HRNet model to extract the contour of the normalized laparoscopic surgical image to obtain the upper and lower edges of the liver in the laparoscopic surgical image, specifically including: The normalized laparoscopic surgical images were used to construct training sets, validation sets, and test sets; Build an HRNet model, train the HRNet model based on the constructed training set, evaluate the performance of the HRNet model based on the constructed validation set, and test the HRNet model training results based on the constructed test set to obtain a trained HRNet model; The trained HRNet model is used to extract the contours of the normalized laparoscopic surgical images to obtain the upper and lower edges of the liver in the laparoscopic surgical images.

[0008] A further improvement of the present invention is that when the HRNet model is trained based on the constructed training set, the loss function used is the focus loss function, and the focus loss function expression is:

[0009] in, is the focal loss function, is the aggregation factor, which is used to control the degree of penalty imposed by the HRNet model on easily classified samples. is the focus factor, which is used to reduce the weight of easy-to-classify samples. is the total number of samples, For the The index of the samples, For the The true labels of samples, For the The predicted probability of a sample, Indicates the confidence of the HRNet model for category 1, For the The logarithm of the predicted probability of samples.

[0010] A further improvement of the present invention is that, when constructing the HRNet model, a channel splicing mechanism is added to improve the HRNet model.

[0011] A further improvement of the present invention is that after improving the HRNet model by adding a channel splicing mechanism, a convolutional block attention module is added.

[0012] A further improvement of the present invention is that the registration calculation of the upper and lower edges of the liver in the laparoscopic surgical image with the upper and lower edges of the liver marked in the constructed preoperative 3D liver model is performed to obtain the registration result, which specifically includes: The RANSAC-PnP algorithm was used to align the upper and lower edges of the liver in the laparoscopic surgical image with the upper and lower edges of the liver marked in the preoperative 3D liver model to obtain the rotation matrix, translation vector, and reprojection error. The registration result is obtained based on the rotation matrix, translation vector and reprojection error.

[0013] In a second aspect, the present invention provides a laparoscopic liver resection surgery registration system based on an artificial intelligence algorithm, comprising a data acquisition module, a data preprocessing module, a data normalization processing module, a contour extraction module, a preoperative model construction module, and a registration module; The data acquisition module is used to acquire laparoscopic surgery images; The data preprocessing module is used to preprocess the obtained laparoscopic surgical images to obtain preprocessed laparoscopic surgical images, wherein the preprocessing includes removing noise and enhancing the laparoscopic surgical images; The data normalization processing module is used to perform normalization processing on the preprocessed laparoscopic surgical images to obtain normalized laparoscopic surgical images; The contour extraction module is used to extract the contour of the normalized laparoscopic surgical image using an artificial intelligence algorithm to obtain the upper edge and lower edge of the liver in the laparoscopic surgical image; The preoperative model construction module is used to construct a preoperative 3D liver model and mark the upper edge and lower edge of the liver in the constructed preoperative 3D liver model; The registration module is used to perform registration calculation on the upper edge and lower edge of the liver in the laparoscopic surgery image with the upper edge and lower edge of the liver marked in the constructed preoperative 3D liver model to obtain a registration result.

[0014] A further improvement of the present invention is that the artificial intelligence algorithm is a HRNet model. A further improvement of the present invention is that the contour extraction module uses the HRNet model to extract the contour of the normalized laparoscopic surgical image to obtain the upper and lower edges of the liver in the laparoscopic surgical image, specifically including: The normalized laparoscopic surgical images were used to construct training sets, validation sets, and test sets; Build an HRNet model, train the HRNet model based on the constructed training set, evaluate the performance of the HRNet model based on the constructed validation set, and test the HRNet model training results based on the constructed test set to obtain a trained HRNet model; The trained HRNet model is used to extract the contours of the normalized laparoscopic surgical images to obtain the upper and lower edges of the liver in the laparoscopic surgical images.

[0015] Compared with the prior art, the present invention has the following beneficial effects: Compared to existing laparoscopic liver resection registration methods, the present invention utilizes an artificial intelligence algorithm to extract the contours of normalized laparoscopic surgical images, obtaining the upper and lower edges of the liver in the laparoscopic images. These contour-extracted upper and lower edges possess specific local features, which can be used in subsequent registration processes to more accurately match different images (laparoscopic surgical images), thereby improving registration accuracy. Furthermore, the present invention performs registration calculations on the upper and lower edges of the liver in the laparoscopic surgical images with the upper and lower edges of the liver marked in a constructed preoperative 3D liver model, achieving precise alignment of the intraoperative image (laparoscopic surgical image) with the three-dimensional model (preoperative 3D liver model). This reduces manual intervention, improves registration accuracy, and effectively addresses the low accuracy of laparoscopic liver resection registration in existing technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a flow chart of the laparoscopic liver resection surgery registration method based on artificial intelligence algorithm of the present invention; Figure 2 Schematic diagram of the laparoscopic liver resection surgery registration system based on artificial intelligence algorithm of the present invention; Figure 3 The present invention provides a laparoscopic surgical image based on an artificial intelligence algorithm and a result after contour extraction of the laparoscopic surgical image; Figure 4 This is a structural diagram of the convolutional block attention module of the present invention; Figure 5 This is the structural diagram of the HRNet model finally adopted by the present invention; Figure 6 Schematic diagram of the preoperative 3D liver model constructed in the present invention. DETAILED DESCRIPTION

[0017] In order to further understand the content of the present invention, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the embodiments are only for explaining the present invention and are not intended to limit it.

[0018] The proposed laparoscopic liver resection registration method, based on an artificial intelligence algorithm, aligns the upper and lower edges of the liver in laparoscopic surgical images with the upper and lower edges of the liver marked in a preoperative 3D liver model. Compared to existing techniques, this method effectively addresses the low accuracy of registration for laparoscopic liver resection.

[0019] Example 1: The flowchart of the laparoscopic liver resection registration method based on artificial intelligence algorithm of the present invention is as follows: Figure 1 As shown, the laparoscopic liver resection surgery registration method based on artificial intelligence algorithm of the present invention includes the following steps: S1. Acquire laparoscopic surgical images; S2. Preprocessing the obtained laparoscopic surgical image to obtain a preprocessed laparoscopic surgical image, wherein the preprocessing includes removing noise and enhancing the laparoscopic surgical image; S3. performing normalization processing on the preprocessed laparoscopic surgical image to obtain a normalized laparoscopic surgical image; S4. Using an artificial intelligence algorithm to extract the contours of the normalized laparoscopic surgical image to obtain the upper and lower edges of the liver in the laparoscopic surgical image; S5. constructing a preoperative 3D liver model, and marking the upper and lower edges of the liver in the constructed preoperative 3D liver model; S6. Perform registration calculation on the upper edge and lower edge of the liver in the laparoscopic surgical image and the upper edge and lower edge of the liver marked in the constructed preoperative 3D liver model to obtain a registration result.

[0020] Example 2: The schematic diagram of the laparoscopic liver resection registration system based on artificial intelligence algorithm of the present invention is shown in FIG. Figure 2 As shown, the laparoscopic liver resection surgery registration system based on artificial intelligence algorithm of the present invention includes a data acquisition module, a data preprocessing module, a data normalization processing module, a contour extraction module, a preoperative model construction module and a registration module.

[0021] The data acquisition module is used to acquire laparoscopic surgery images.

[0022] The data preprocessing module is used to preprocess the obtained laparoscopic surgical images to obtain preprocessed laparoscopic surgical images, wherein the preprocessing includes noise removal and laparoscopic surgical image enhancement.

[0023] The data normalization processing module is used to perform normalization processing on the preprocessed laparoscopic surgical images to obtain normalized laparoscopic surgical images.

[0024] The contour extraction module is used to extract the contour of the normalized laparoscopic surgical image using an artificial intelligence algorithm to obtain the upper and lower edges of the liver in the laparoscopic surgical image.

[0025] The preoperative model construction module is used to construct a preoperative 3D liver model and mark the upper edge and lower edge of the liver in the constructed preoperative 3D liver model.

[0026] The registration module is used to perform registration calculation on the upper and lower edges of the liver in the laparoscopic surgical image with the upper and lower edges of the liver marked in the constructed preoperative 3D liver model to obtain a registration result.

[0027] Example 3: The laparoscopic liver resection registration method based on artificial intelligence algorithm of the present invention comprises the following steps: S1. Acquire laparoscopic surgical images.

[0028] S2. Preprocess the obtained laparoscopic surgical image to obtain a preprocessed laparoscopic surgical image, where the preprocessing includes removing noise and enhancing the laparoscopic surgical image.

[0029] In this step, Gaussian filtering or bilateral filtering is used to remove noise. Specifically, the laparoscopic surgical image is enhanced by adjusting the brightness and saturation of the HSV (Hue, Saturation, Value) or CIELAB (International Commission on Illumination LAB color space).

[0030] S3. Normalize the preprocessed laparoscopic surgical image to obtain a normalized laparoscopic surgical image.

[0031] The calculation formula for normalization in this step is:

[0032] in, is the pre-processed laparoscopic surgical image, is the normalized laparoscopic surgical image, is the minimum value of the preprocessed laparoscopic surgery image, Usually the value is 0. is the maximum value of the preprocessed laparoscopic surgery image, The value is usually 255.

[0033] S4. Use artificial intelligence algorithm to extract the contour of the normalized laparoscopic surgical image to obtain the upper and lower edges of the liver in the laparoscopic surgical image.

[0034] This step specifically uses the HRNet model to extract the contours of the normalized laparoscopic surgical image to obtain the upper and lower edges of the liver in the laparoscopic surgical image.

[0035] This step uses the HRNet model to extract the contours of the normalized laparoscopic surgical image to obtain the upper and lower edges of the liver in the laparoscopic surgical image. Specifically, it includes: A. Use normalized laparoscopic surgical images to construct training sets, validation sets, and test sets; B. Build an HRNet model, train the HRNet model based on the constructed training set, evaluate the performance of the HRNet model based on the constructed validation set, and test the HRNet model training results based on the constructed test set to obtain a trained HRNet model; C. Use the trained HRNet model to extract the contours of the normalized laparoscopic surgical image to obtain the upper and lower edges of the liver in the laparoscopic surgical image.

[0036] In step A, there are 921 training images (training set), 122 validation images (validation set) and 109 test images (test set). The pixel size of each image is 1920×1080. During the training process, the size of each image is fixed to 1024×1024. The laparoscopic surgical images and the results after the contour extraction of the laparoscopic surgical images are shown in the figure. Figure 3 As shown, Figure 3 a is a laparoscopic surgery image, Figure 3 b is the result of contour extraction of laparoscopic surgery images. Figure 3 The green part in b is the upper edge of the liver in the laparoscopic surgery image (also called the upper margin). Figure 3 The red part in b is the lower edge of the liver (also called the lower margin) in the laparoscopic surgery image, and the blue part is the resected falciform ligament. Figure 3 The black part in b is the background.

[0037] When constructing the HRNet model in step B, a channel splicing mechanism is added to improve the HRNet model.

[0038] The performance indicators of the HRNet model based on the constructed validation set in step B include precision, recall, F1 score and mean intersection over union (mIoU).

[0039] This example compares the evaluation index results of the existing four models and the HRNet model. The comparison results are shown in Table 1.

[0040] Table 1 Comparison of evaluation index results

[0041] The data in Table 1 show that the four indicators of the HRNet model are all higher than those of the other four models. Therefore, this embodiment chooses to use the HRNet model to perform contour extraction on the normalized laparoscopic surgical images to obtain the upper and lower edges of the liver in the laparoscopic surgical images.

[0042] When training the HRNet model based on the constructed training set in step B, the loss function used is the focus loss function, and the focus loss function expression is:

[0043] in, is the focal loss function, is the aggregation factor, which is used to control the degree of penalty imposed by the HRNet model on easily classified samples. Usually the value is 2. is the focus factor, which is used to reduce the weight of easy-to-classify samples. is the total number of samples, For the The index of the samples, For the The true labels of samples, For the The predicted probability of a sample, Indicates the confidence of the HRNet model for category 1, For the The logarithm of the predicted probability of samples.

[0044] After improving the HRNet model by adding the channel splicing mechanism in step B, the Convolutional Block Attention Module (CBAM) is added. The following is a detailed description of the Convolutional Block Attention Module: CBAM is a lightweight and efficient attention mechanism commonly used in convolutional neural networks (CNNs) to improve feature representation. By introducing an attention mechanism, CBAM enables CNNs to focus more on key spatial and channel information.

[0045] The core idea of ​​CBAM is to model attention in the channel dimension and spatial dimension of the feature map. CBAM mainly consists of a channel attention module and a spatial attention module. Channel attention allows the convolutional neural network to pay more attention to which channels contain more important information. Channel attention uses global average pooling and global maximum pooling to extract channel dimension features, representing different statistical information of the feature map respectively. The structure diagram of the convolutional block attention module is shown in the figure below. Figure 4 shown.

[0046] This example compares the evaluation index results after HRNet model optimization, and the comparison results are shown in Table 2.

[0047] Table 2 Comparison of evaluation index results

[0048] In Table 2, Optimization 1 uses the focal loss function when training the HRNet model based on the constructed training set. Optimization 2 improves the HRNet model by adding a channel splicing mechanism. Optimization 3 improves the HRNet model by adding a convolutional block attention module. When only Optimization 2 and Optimization 3 are used, the cross-entropy loss function is used when training the HRNet model based on the constructed training set.

[0049] As can be seen from Table 2, the evaluation indicators (precision, recall, F1 score and average intersection-over-union) of the optimized HRNet model are improved overall compared with the evaluation indicators (precision, recall, F1 score and average intersection-over-union) of the HRNet model in Table 1.

[0050] The structural diagram of the HRNet model finally adopted by the present invention is as follows Figure 5 As shown, the process of obtaining the upper and lower edges of the liver in laparoscopic surgery images is described in detail below: First, convolution and downsampling operations are performed on the normalized laparoscopic surgical images to obtain four parallel branches, each with a feature map with a different resolution. Then, in the first three branches, channel splicing (also called adding a channel splicing mechanism) is performed on the feature maps with the same resolution to obtain three feature maps with different resolutions. Finally, the feature maps with three different resolutions and the feature map of the fourth branch are simultaneously processed using a convolutional block attention module (also called a CBAM attention module). After being processed simultaneously using the convolutional block attention module, an upsampling operation is performed to obtain the upper and lower edges of the liver in the laparoscopic surgical image.

[0051] S5. Construct a preoperative 3D liver model, and mark the upper and lower edges of the liver in the constructed preoperative 3D liver model.

[0052] A preoperative 3D liver model was constructed, and the upper and lower edges of the liver were marked in the constructed preoperative 3D liver model. Figure 6 shown.

[0053] S6. Perform registration calculation on the upper edge and lower edge of the liver in the laparoscopic surgical image and the upper edge and lower edge of the liver marked in the constructed preoperative 3D liver model to obtain a registration result.

[0054] The upper edge and the lower edge of the liver in the laparoscopic surgery image are matched with the upper edge and the lower edge of the liver marked in the constructed preoperative 3D liver model to obtain a matching result, specifically including: a. The RANSAC-PnP algorithm is used to match the upper edge and the lower edge of the liver in the laparoscopic surgery image with the upper edge and the lower edge of the liver marked in the constructed preoperative 3D liver model to obtain a rotation matrix, a translation vector and a re-projection error. The specific process of obtaining the rotation matrix in this step is as follows: First, define three unit vectors , and , , and The calculation formula is:

[0055]

[0056]

[0057] Among them, , and are unit vectors, x represents the right direction of the camera, represents the upward direction of the camera, represents the line-of-sight direction of the camera (from the focal point to the camera), is the direction vector on the camera, is the camera position, is the focal point position.

[0058] Finally, the rotation matrix obtained through the unit vectors , and is:

[0059] Among them, is the rotation matrix, is the coordinate value of the vector, is the coordinate value of the vector, is the coordinate value of the vector, is the coordinate value of the vector, is the coordinate value of the vector, is the coordinate value of the vector, is the coordinate value of the vector, Coordinate values, are of the vector Coordinate values, are of the vector Coordinate values, are of the vector Coordinate values, are of the vector Coordinate values.

[0060] The calculation formula of the translation vector in this step is:

[0061] wherein, is the translation vector, is the rotation matrix, is the camera position.

[0062] wherein, the above-mentioned camera position , focal point position and direction vector on the camera are obtained through the rendering library vtkcamera in the data visualization development package vtk.

[0063] The calculation formula of the re-projection error in this step is:

[0064] wherein, is the re-projection error, is the number of points participating in registration, is the re-projection error of the th pixel point, is the point of the laparoscopic surgery image pixel, is the pixel coordinate calculated through the rotation matrix is the row coordinate of the pixel point, represents and Euclidean distance.

[0065] b. Based on the rotation matrix, the translation vector and the re-projection error, the registration result is obtained.

[0066] When the rotation matrix, the translation vector and the re-projection error all meet the set conditions in this step, it means that the upper edge and the lower edge of the liver of the laparoscopic surgery image can be registered and calculated with the upper edge and the lower edge of the liver marked in the constructed preoperative 3D liver model.

[0067] ​Specifically, when the rotation matrix and the translation vector can make the upper and lower edges of the liver in the laparoscopic surgical image coincide with the upper and lower edges of the liver marked in the constructed preoperative 3D liver model within a set range (the set range is adjusted according to actual needs), it means that the upper and lower edges of the liver in the laparoscopic surgical image can be aligned with the upper and lower edges of the liver marked in the constructed preoperative 3D liver model.

[0068] Specifically, when the reprojection error meets the set error threshold (the error threshold in this step can be adjusted according to actual needs, and the error threshold in this step is generally set to 4-5 mm), it means that the upper and lower edges of the liver in the laparoscopic surgical image can be aligned with the upper and lower edges of the liver marked in the constructed preoperative 3D liver model.

[0069] The registration result is finally obtained according to the registration calculation process.

[0070] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A laparoscopic liver resection registration method based on artificial intelligence algorithm, characterized in that: The following steps are involved: Acquire laparoscopic surgical images; Preprocessing the obtained laparoscopic surgical image to obtain a preprocessed laparoscopic surgical image, wherein the preprocessing includes removing noise and enhancing the laparoscopic surgical image; performing normalization processing on the preprocessed laparoscopic surgical image to obtain a normalized laparoscopic surgical image; Artificial intelligence algorithms were used to extract the contours of the normalized laparoscopic surgical images to obtain the upper and lower edges of the liver in the laparoscopic surgical images. Constructing a preoperative 3D liver model and marking the upper and lower edges of the liver in the constructed preoperative 3D liver model; The upper and lower edges of the liver in the laparoscopic surgical image are registered with the upper and lower edges of the liver marked in the constructed preoperative 3D liver model to obtain a registration result.

2. The registration method for laparoscopic liver resection surgery based on artificial intelligence algorithm according to claim 1, characterized in that: The artificial intelligence algorithm is the HRNet model.

3. The registration method for laparoscopic liver resection surgery based on artificial intelligence algorithm according to claim 2, characterized in that: The HRNet model is used to extract the contours of the normalized laparoscopic surgical images to obtain the upper and lower edges of the liver in the laparoscopic surgical images, including: The normalized laparoscopic surgical images were used to construct training sets, validation sets, and test sets; Build an HRNet model, train the HRNet model based on the constructed training set, evaluate the performance of the HRNet model based on the constructed validation set, and test the HRNet model training results based on the constructed test set to obtain a trained HRNet model; The trained HRNet model is used to extract the contours of the normalized laparoscopic surgical images to obtain the upper and lower edges of the liver in the laparoscopic surgical images.

4. The registration method for laparoscopic liver resection surgery based on artificial intelligence algorithm according to claim 3, characterized in that: When training the HRNet model based on the constructed training set, the loss function used is the focus loss function, and the focus loss function expression is: in, is the focal loss function, is the aggregation factor, which is used to control the degree of penalty imposed by the HRNet model on easily classified samples. is the focus factor, which is used to reduce the weight of easy-to-classify samples. is the total number of samples, For the The index of the samples, For the The true labels of samples, For the The predicted probability of a sample, Indicates the confidence of the HRNet model for category 1, For the The logarithm of the predicted probability of samples.

5. The registration method for laparoscopic liver resection surgery based on artificial intelligence algorithm according to claim 3, characterized in that: When building the HRNet model, a channel splicing mechanism is added to improve the HRNet model.

6. The registration method for laparoscopic liver resection surgery based on artificial intelligence algorithm according to claim 5, characterized in that: After improving the HRNet model by adding the channel splicing mechanism, the convolutional block attention module is added.

7. The registration method for laparoscopic liver resection surgery based on artificial intelligence algorithm according to claim 1, characterized in that: The registration calculation of the upper edge and lower edge of the liver in the laparoscopic surgery image with the upper edge and lower edge of the liver marked in the constructed preoperative 3D liver model to obtain the registration result specifically includes: The RANSAC-PnP algorithm was used to align the upper and lower edges of the liver in the laparoscopic surgical image with the upper and lower edges of the liver marked in the preoperative 3D liver model to obtain the rotation matrix, translation vector, and reprojection error. The registration result is obtained based on the rotation matrix, translation vector and reprojection error.

8. A laparoscopic liver resection registration system based on artificial intelligence algorithm, characterized by: It includes data acquisition module, data preprocessing module, data normalization processing module, contour extraction module, preoperative model building module and registration module; The data acquisition module is used to acquire laparoscopic surgery images; The data preprocessing module is used to preprocess the obtained laparoscopic surgical images to obtain preprocessed laparoscopic surgical images, wherein the preprocessing includes removing noise and enhancing the laparoscopic surgical images; The data normalization processing module is used to perform normalization processing on the preprocessed laparoscopic surgical images to obtain normalized laparoscopic surgical images; The contour extraction module is used to extract the contour of the normalized laparoscopic surgical image using an artificial intelligence algorithm to obtain the upper edge and lower edge of the liver in the laparoscopic surgical image; The preoperative model construction module is used to construct a preoperative 3D liver model and mark the upper edge and lower edge of the liver in the constructed preoperative 3D liver model; The registration module is used to perform registration calculation on the upper edge and lower edge of the liver in the laparoscopic surgery image with the upper edge and lower edge of the liver marked in the constructed preoperative 3D liver model to obtain a registration result.

9. The laparoscopic liver resection registration system based on artificial intelligence algorithm according to claim 8, characterized in that: The artificial intelligence algorithm is the HRNet model.

10. The laparoscopic liver resection registration system based on artificial intelligence algorithm according to claim 9, characterized in that: The contour extraction module uses the HRNet model to extract the contours of the normalized laparoscopic surgical images to obtain the upper and lower edges of the liver in the laparoscopic surgical images. Specifically, it includes: The normalized laparoscopic surgical images were used to construct training sets, validation sets, and test sets; Build an HRNet model, train the HRNet model based on the constructed training set, evaluate the performance of the HRNet model based on the constructed validation set, and test the HRNet model training results based on the constructed test set to obtain a trained HRNet model; The trained HRNet model is used to extract the contours of the normalized laparoscopic surgical images to obtain the upper and lower edges of the liver in the laparoscopic surgical images.

Images