Gynecological operation navigation method and system based on minimally invasive technology

Abstract

The invention relates to the technical field of image recognition, in particular to a gynecological operation navigation method and system based on the minimally invasive technology, and the method comprises the steps: carrying out the feature recognition operation of a three-dimensional focus region structure diagram, and obtaining a focus surrounding tissue feature set; performing minimally invasive surgery scheme identification operation on the focus surrounding tissue feature set to obtain a predicted channel number; performing minimally invasive surgery path recognition operation based on the predicted channel number on the tissue feature set around the focus to obtain a minimally invasive path set, and acquiring necessary avoidance tissue sequences in a preset range region on each minimally invasive path in the minimally invasive path set to obtain a necessary avoidance tissue sequence set; according to the necessary avoiding tissue sequence set, the real-time instrument position information, the minimally invasive path set and the path distribution purpose type set, cutting operation is conducted on the target treatment area, and a minimally invasive channel set is obtained. The navigation accuracy of the minimally invasive surgery can be improved, and the safety of the minimally invasive surgery is improved.

Description

Technical Field

[0001] The present invention relates to the field of image recognition technology, and in particular to a gynecological surgery navigation method and system based on minimally invasive technology. Background Art

[0002] In recent years, with the rapid development of minimally invasive gynecological surgery technology, its application in clinical practice has become increasingly widespread.

[0003] However, traditional minimally invasive surgery still faces numerous challenges: The surgeon relies on experience to determine the spatial relationship between the instrument and tissue, limiting operational precision. Complex cases are prone to accidental damage to vital blood vessels and nerves due to anatomical variations or limited visual field. The incidence of postoperative complications (such as bleeding and infection) is closely related to surgical precision. Existing image navigation technology suffers from a delay in the fusion of preoperative and intraoperative data. If the preoperative minimally invasive path positioning accuracy is poor, it can easily lead to increased difficulty in the intraoperative cutting and treatment process. Summary of the Invention

[0004] The present invention provides a gynecological surgery navigation method based on minimally invasive technology, the main purpose of which is to improve the accuracy of minimally invasive surgery navigation and enhance the safety of minimally invasive surgery.

[0005] To achieve the above objectives, the present invention provides a gynecological surgery navigation method based on minimally invasive technology, comprising:

[0006] Obtaining a CT scan image and an MRI image of a preset target treatment area, performing three-dimensional modeling on the CT scan image and the MRI image to obtain a three-dimensional lesion area structure diagram;

[0007] Performing a feature recognition operation on the three-dimensional lesion region structure image to obtain a set of tissue features surrounding the lesion;

[0008] Performing a minimally invasive surgery plan recognition operation on the set of tissue features surrounding the lesion to obtain a predicted channel number;

[0009] performing a minimally invasive surgical path identification operation based on the predicted number of channels on the set of tissue features surrounding the lesion to obtain a minimally invasive path set, and obtaining a necessary tissue avoidance sequence within a preset range area on each minimally invasive path in the minimally invasive path set to obtain a necessary tissue avoidance sequence set;

[0010] performing a path usage allocation identification operation on each minimally invasive path in the minimally invasive path set according to the necessary avoidance tissue sequence set to obtain a path allocation usage type set;

[0011] Perform electromagnetic positioning on the pre-built cutting and puncturing instrument to obtain real-time instrument position information, assign a usage type set according to the path, obtain the wound range of each minimally invasive path in the minimally invasive path set, and obtain wound range information;

[0012] The cutting and puncturing instrument is used to perform a cutting operation on the target treatment area according to the real-time instrument position information, the minimally invasive path set and the wound range information to obtain a minimally invasive channel set.

[0013] Optionally, obtaining a CT scan image and an MRI image of a preset target treatment area includes:

[0014] Using a pre-built CT scanning device, a CT scan is performed on a preset target treatment area to obtain an initial CT scan image, and using a pre-built nuclear magnetic resonance device, a nuclear magnetic resonance scan is performed on the target treatment area to obtain an initial MRI image;

[0015] performing Gaussian filtering on the initial CT scan image to obtain a noise-reduced CT image, performing frequency domain processing based on Fourier transform on the noise-reduced CT image to obtain an artifact-corrected CT image, and performing CT value normalization on the artifact-corrected CT image to obtain a standard CT image;

[0016] Performing anisotropic denoising on the initial MRI image to obtain a denoised MRI image, and performing spatial normalization on the denoised MRI image according to a preset standard template to obtain a standard MRI image;

[0017] The standard CT image and the standard MRI image are aligned using a pre-built mutual information algorithm to obtain a CT scan image and an MRI image.

[0018] Optionally, performing three-dimensional modeling on the CT scan image and the MRI image to obtain a three-dimensional lesion area structure diagram includes:

[0019] Using a pre-built image recognition model, performing a feature extraction operation on the CT scan image to obtain a CT image feature set, and performing a feature extraction operation on the MRI image to obtain an MRI image feature set;

[0020] performing feature importance analysis on the CT image feature set according to a preset target recognition type set to obtain a first confidence sequence;

[0021] Performing feature importance analysis on the MRI image feature set to obtain a second confidence sequence;

[0022] Obtaining, based on the first confidence sequence and the second confidence sequence, a feature allocation weight for each target recognition type in the target recognition type set to obtain a weight configuration set;

[0023] performing a fully connected recognition operation based on the target recognition type set on the CT image feature set and the MRI image feature set according to the weight configuration set to obtain a target recognition result set;

[0024] The positional relationship of each target recognition result in the target recognition result set is obtained to obtain position distribution information, and the target recognition result set is arranged in three dimensions according to the position distribution information to obtain a three-dimensional lesion area structure map.

[0025] Optionally, performing a feature recognition operation on the three-dimensional lesion region structure image to obtain a set of tissue features surrounding the lesion includes:

[0026] Using a pre-constructed convolution kernel set, a traversal convolution operation is performed on the three-dimensional lesion area structure map to obtain a convolution matrix set;

[0027] Performing an average pooling operation on the convolution matrix set to obtain a pooling matrix set;

[0028] A flattening operation is performed on each pooling matrix in the pooling matrix set to obtain a set of perilesion tissue features.

[0029] Optionally, performing a minimally invasive surgery plan identification operation on the set of tissue features surrounding the lesion to obtain a predicted number of channels includes:

[0030] Performing lesion type identification on the set of tissue features surrounding the lesion to obtain the lesion type;

[0031] Using a pre-built surgical standardized channel number table, query the channel number corresponding to the lesion type to obtain an initial channel number;

[0032] Performing lesion complexity identification on the set of tissue features surrounding the lesion to obtain a complexity score;

[0033] According to the complexity score, the initial number of channels is adaptively adjusted to obtain a predicted number of channels.

[0034] Optionally, performing a minimally invasive surgical path identification operation based on the predicted number of channels on the tissue feature set surrounding the lesion to obtain a minimally invasive path set includes:

[0035] Obtaining the location of the lesion based on the set of tissue features surrounding the lesion, and obtaining the shortest penetration route between a preset external region and the lesion location;

[0036] Obtaining target recognition results of the shortest penetration route within a preset wound range from the target recognition result set to obtain a damaged target set;

[0037] Using a pre-built avoidance weight table, querying the avoidance score of each damaged target in the damaged target set to obtain an avoidance score set, and obtaining a total avoidance score of the avoidance score set;

[0038] Determining whether the total avoidance score is greater than a preset cutting threshold;

[0039] When the total avoidance score is greater than or equal to the cutting threshold, the shortest penetration path is determined to be unqualified, and the shortest penetration path is fine-tuned according to the pre-built genetic algorithm and the damage target set to obtain an updated penetration path. The updated penetration path is used to replace the shortest penetration path, and the process returns to the above step of obtaining the target recognition result of the shortest penetration path within the preset wound range from the target recognition result set;

[0040] When the total avoidance score is less than the cutting threshold, the shortest penetration path is determined to be qualified, and the shortest penetration path is sent to a pre-built candidate path database to obtain a qualified minimally invasive path;

[0041] A minimally invasive pathway set is obtained from the candidate pathway database.

[0042] Optionally, obtaining a minimally invasive pathway set from the candidate pathway database includes:

[0043] Obtaining minimally invasive paths in the candidate path database to obtain stored paths and the number of paths;

[0044] Determining whether the number of paths is less than the predicted number of channels;

[0045] When the number of paths is less than the predicted number of channels, performing a minimally invasive surgical path identification operation based on the predicted number of channels on the set of tissue features surrounding the lesion, obtaining a minimally invasive path other than the stored path;

[0046] When the number of paths is equal to the number of predicted channels, the stored paths are output to obtain a minimally invasive path set.

[0047] Optionally, after obtaining the minimally invasive channel set, the method further includes:

[0048] Acquire a pre-built surgical instrument set, perform electromagnetic positioning on the surgical instrument set, and obtain a surgical instrument position set;

[0049] Obtaining the instrument type of each surgical instrument in the surgical instrument set to obtain an instrument type set;

[0050] According to the instrument type set, the minimally invasive path set and the surgical instrument position set, each surgical instrument in the surgical instrument set is passed through a corresponding minimally invasive channel in the minimally invasive channel set.

[0051] To achieve the above objectives, the present invention further provides a gynecological surgery navigation system based on minimally invasive technology, comprising:

[0052] An information acquisition module is configured to acquire a CT scan image and an MRI image of a preset target treatment area, perform three-dimensional modeling on the CT scan image and the MRI image to obtain a three-dimensional lesion area structure diagram, and perform feature recognition on the three-dimensional lesion area structure diagram to obtain a set of tissue features surrounding the lesion;

[0053] a minimally invasive path positioning and allocation module, configured to perform a minimally invasive surgical plan identification operation on the set of tissue features surrounding the lesion to obtain a predicted number of channels, perform a minimally invasive surgical path identification operation based on the predicted number of channels on the set of tissue features surrounding the lesion to obtain a minimally invasive path set, obtain a necessary tissue avoidance sequence within a preset range area on each minimally invasive path in the minimally invasive path set to obtain a necessary tissue avoidance sequence set, and perform a path usage allocation identification operation on each minimally invasive path in the minimally invasive path set based on the necessary tissue avoidance sequence set to obtain a path allocation usage type set;

[0054] The minimally invasive channel guided puncture module is used to electromagnetically locate the pre-built cutting and puncture instrument to obtain real-time instrument position information, allocate a use type set according to the path, obtain the wound range of each minimally invasive path in the minimally invasive path set, obtain wound range information, and use the cutting and puncture instrument to perform a cutting operation on the target treatment area according to the real-time instrument position information, the minimally invasive path set and the wound range information to obtain a minimally invasive channel set.

[0055] Optionally, performing a minimally invasive surgical path identification operation based on the predicted number of channels on the tissue feature set surrounding the lesion to obtain a minimally invasive path set includes:

[0056] Obtaining the location of the lesion based on the set of tissue features surrounding the lesion, and obtaining the shortest penetration route between a preset external region and the lesion location;

[0057] Obtaining target recognition results of the shortest penetration route within a preset wound range from the target recognition result set to obtain a damaged target set;

[0058] Using a pre-built avoidance weight table, querying the avoidance score of each damaged target in the damaged target set to obtain an avoidance score set, and obtaining a total avoidance score of the avoidance score set;

[0059] Determining whether the total avoidance score is greater than a preset cutting threshold;

[0060] When the total avoidance score is greater than or equal to the cutting threshold, the shortest penetration path is determined to be unqualified, and the shortest penetration path is fine-tuned according to the pre-built genetic algorithm and the damage target set to obtain an updated penetration path. The updated penetration path is used to replace the shortest penetration path, and the process returns to the above step of obtaining the target recognition result of the shortest penetration path within the preset wound range from the target recognition result set;

[0061] When the total avoidance score is less than the cutting threshold, the shortest penetration path is determined to be qualified, and the shortest penetration path is sent to a pre-built candidate path database to obtain a qualified minimally invasive path;

[0062] A minimally invasive pathway set is obtained from the candidate pathway database.

[0063] In order to solve the above problem, the present invention further provides an electronic device, comprising:

[0064] a memory storing at least one instruction;

[0065] The processor executes the instructions stored in the memory to implement the above-mentioned gynecological surgery navigation method based on minimally invasive technology.

[0066] In order to solve the above problems, the present invention also provides a computer-readable storage medium, which stores at least one instruction, and the at least one instruction is executed by a processor in an electronic device to implement the above-mentioned gynecological surgery navigation method based on minimally invasive technology.

[0067] The present invention solves the problems described in the background technology. The present invention uses multi-source data from CT scans and MRI scans to identify the structures surrounding the lesions in the target treatment area and obtain a three-dimensional lesion area structure map. Among them, CT scans focus more on the identification of bones and blood vessel distribution, while MRI scans focus more on the identification of lesions, nerves, and soft tissues. Through the collaborative identification of the two, the clarity and accuracy of the three-dimensional lesion area structure map are greatly improved. Then, based on the set of tissue features surrounding the lesions in the three-dimensional lesion area structure map, the present invention identifies multiple channels at a time, and then, based on the distribution of different tissue features involved in each minimally invasive path, assigns different minimally invasive paths to the task requirements of different medical devices, obtains a minimally invasive path set and a path allocation use type set, thereby effectively providing a route for the subsequent puncture process of the cutting and puncturing instrument, and then guiding the cutting and puncturing instrument to puncture the lesion area. Therefore, the present invention can improve the accuracy of minimally invasive surgical navigation and improve the safety of minimally invasive surgery. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 A schematic diagram of a process flow of a gynecological surgery navigation method based on minimally invasive technology provided by one embodiment of the present invention;

[0069] Figure 2 A functional module diagram of a gynecological surgery navigation system based on minimally invasive technology provided by one embodiment of the present invention;

[0070] Figure 3 A schematic structural diagram of an electronic device for implementing the minimally invasive gynecological surgery navigation method provided in one embodiment of the present invention.

[0071] Description of reference numerals:

[0072] 1. Electronic device; 10. Processor; 11. Memory; 12. Bus.

[0073] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0074] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0075] The embodiments of the present application provide a minimally invasive technology-based gynecological surgery navigation method. The execution subject of the minimally invasive technology-based gynecological surgery navigation method includes, but is not limited to, at least one of electronic devices such as a server and a terminal that can be configured to execute the method provided by the embodiments of the present application. In other words, the minimally invasive technology-based gynecological surgery navigation method can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0076] Reference Figure 1 FIG. 1 is a flow chart of a minimally invasive gynecological surgery navigation method according to an embodiment of the present invention. In this embodiment, the minimally invasive gynecological surgery navigation method includes:

[0077] S1. Obtain a CT scan image and an MRI image of a preset target treatment area, perform three-dimensional modeling on the CT scan image and the MRI image, and obtain a three-dimensional lesion area structure diagram.

[0078] The target treatment area refers to the area where the lesion is located as assessed by the doctor.

[0079] The CT scan image is a cross-sectional or three-dimensional medical image generated by computer processing after layered scanning of the human body by an X-ray beam.

[0080] The MRI image is a medical image generated based on the principle of nuclear magnetic resonance, which can non-invasively display the internal structure and pathological information of the human body.

[0081] The three-dimensional modeling refers to the process of arranging the position distribution of each tissue in the image according to the three-dimensional space. The three-dimensional lesion area structure diagram refers to the three-dimensional modeling results of each tissue shown in the CT scan image and MRI image.

[0082] In detail, in an embodiment of the present invention, obtaining a CT scan image and an MRI image of a preset target treatment area includes:

[0083] Using a pre-built CT scanning device, a CT scan is performed on a preset target treatment area to obtain an initial CT scan image, and using a pre-built nuclear magnetic resonance device, a nuclear magnetic resonance scan is performed on the target treatment area to obtain an initial MRI image;

[0084] performing Gaussian filtering on the initial CT scan image to obtain a noise-reduced CT image, performing frequency domain processing based on Fourier transform on the noise-reduced CT image to obtain an artifact-corrected CT image, and performing CT value normalization on the artifact-corrected CT image to obtain a standard CT image;

[0085] Performing anisotropic denoising on the initial MRI image to obtain a denoised MRI image, and performing spatial normalization on the denoised MRI image according to a preset standard template to obtain a standard MRI image;

[0086] The standard CT image and the standard MRI image are aligned using a pre-built mutual information algorithm to obtain a CT scan image and an MRI image.

[0087] The CT scanner is an X-ray-based medical imaging device that generates cross-sectional or three-dimensional images of the human body's internal structures through tomography and computer reconstruction. The CT scan refers to the process performed by the CT scanner. The initial CT scan image refers to the result directly output from the CT scanner.

[0088] The MRI device is a non-invasive medical imaging device based on the principles of nuclear magnetic resonance (NMR). It uses a strong magnetic field and radiofrequency waves to generate high-resolution images of the human body's internal structures. The MRI scan refers to the process performed by the NMR device on the target treatment area. The initial MRI image refers to the direct output of the NMR device.

[0089] The Gaussian filter is a linear image smoothing technique based on the normal distribution function, which is mainly used to remove noise and blur image details. The de-noised CT image refers to the result of Gaussian filter processing of the initial CT scan image.

[0090] The Fourier transform-based frequency domain processing refers to the process of performing frequency domain analysis using Fourier transform to locate artifact features and filter them out, and then restoring valid information through inverse transform. The artifact-corrected CT image refers to the frequency domain processing result of the noise-reduced CT image.

[0091] CT value standardization refers to a series of technical processes and standardized measures to ensure the comparability and diagnostic reliability of CT values ​​(Houinf units, HU) obtained from different equipment and under different scanning conditions. These processes include dark field correction, air correction, and scatter correction, which are not detailed here. The standard CT image refers to an artifact-corrected CT image after CT value standardization.

[0092] Anisotropic denoising is a nonlinear filtering technique based on edge preservation. It is used to smooth noise while preserving image structural features by dynamically adjusting the diffusion coefficient. Its core concept is derived from thermal diffusion theory. It treats the image as a thermal field and pixel values ​​as temperatures, achieving selective smoothing by controlling the heat conduction rate in different directions. The denoised MRI image is the result of anisotropic denoising of the initial MRI image.

[0093] The standard template is configured as the MNI space.

[0094] The mutual information algorithm is an information theory metric used to measure the degree of interdependence between two random variables and is widely used in fields such as feature selection, data mining, and machine learning. Alignment is the execution of the mutual information algorithm. The CT scan image and MRI image are standard CT images and standard MRI images that have undergone information alignment.

[0095] Specifically, the present invention utilizes SimpleITK or PyDicom libraries to implement batch processing of Gaussian filtering and Fourier transforms, obtaining a denoised CT image and an artifact-corrected CT image. CT values ​​(Hounsfield Units) are then mapped to a fixed range (e.g., [-1000, 400] corresponding to air to soft tissue) and CT value normalization is performed to obtain a standard CT image.

[0096] Furthermore, the present invention uses a Perona-Malik filter or a Weickert filter to perform an anisotropic denoising operation on the initial MRI image to obtain a denoised MRI image, and uses an FSL or SPM tool to perform a spatial normalization operation on the denoised MRI image to obtain a standard MRI image.

[0097] Specifically, in the embodiment of the present invention, ITK or 3D Slicer is used to execute a mutual information algorithm to achieve rigid / non-rigid registration, so that the error between the aligned CT and MRI images is less than 2 mm.

[0098] In detail, in an embodiment of the present invention, performing three-dimensional modeling on the CT scan image and the MRI image to obtain a three-dimensional lesion area structure diagram includes:

[0099] Using a pre-built image recognition model, performing a feature extraction operation on the CT scan image to obtain a CT image feature set, and performing a feature extraction operation on the MRI image to obtain an MRI image feature set;

[0100] performing feature importance analysis on the CT image feature set according to a preset target recognition type set to obtain a first confidence sequence;

[0101] Performing feature importance analysis on the MRI image feature set to obtain a second confidence sequence;

[0102] Obtaining, based on the first confidence sequence and the second confidence sequence, a feature allocation weight for each target recognition type in the target recognition type set to obtain a weight configuration set;

[0103] performing a fully connected recognition operation based on the target recognition type set on the CT image feature set and the MRI image feature set according to the weight configuration set to obtain a target recognition result set;

[0104] The positional relationship of each target recognition result in the target recognition result set is obtained to obtain position distribution information, and the target recognition result set is arranged in three dimensions according to the position distribution information to obtain a three-dimensional lesion area structure map.

[0105] The image recognition model is a CNN neural network used for image feature extraction and can be directly accessed and deployed from the internet. Feature extraction, a common neural network function, converts input content into vector information and then extracts key information from that vector information. The CT image feature set is the feature extraction result of a CT scan image.

[0106] The principle of performing feature extraction on the MRI image is the same as that of performing feature extraction on the CT scan image, with the object being the MRI image. The MRI image feature set refers to the feature extraction result of the MRI image.

[0107] The target identification type set refers to tissue types that medical experts believe cannot be easily damaged, such as important nerves, blood vessels with a diameter greater than 3 mm, bones, some flesh tissues, cartilage, etc.

[0108] The feature importance analysis refers to analyzing the confidence of each feature for identifying each target recognition type. The first confidence sequence refers to the distribution of the confidence of the CT image feature set for the target recognition types in the target recognition type set. The second confidence sequence refers to the distribution of the confidence of the MRI image feature set for the target recognition types in the target recognition type set.

[0109] The feature allocation weight is a ratio of a CT image feature set to an MRI image feature set for target recognition type. The weight configuration set refers to a set of feature allocation weights.

[0110] The fully connected recognition operation based on the target recognition type set refers to performing a weighted calculation on the CT image features and MRI image features corresponding to the same tissue during the fully connected recognition process, followed by a fully connected operation on each weighted vector. This fully connected operation is a target recognition type recognition operation and is a common computational process in neural networks, which will not be described in detail here. The target recognition result set refers to the output of the image recognition model, which includes distribution information of various tissue structures.

[0111] The positional relationship refers to the relative position of each target recognition result. The position distribution information refers to the spatial distribution information of each target recognition result after combining the positional relationships. The three-dimensional arrangement refers to the process of three-dimensionally composing the target recognition result set based on the position distribution information.

[0112] Specifically, the present invention directly calls a pre-built image recognition model to perform feature extraction operations on CT scan images and MRI images, and obtains a CT image feature set and an MRI image feature set, respectively.

[0113] Specifically, in an embodiment of the present invention, CT images are used to observe the contrast of various tissues (e.g., lung windows display lung nodules, and bone windows observe bone destruction), and have a high degree of confidence in observing blood vessels, bones, tumors, etc., while MRI, with its high soft tissue resolution and multi-parameter imaging (e.g., T1WI, T2WI, DWI), can clearly distinguish muscles, ligaments, nerve bundles, and tumor boundaries, and therefore has a high degree of confidence in observing soft tissues, nerves, etc. Therefore, the present invention performs a feature importance analysis operation based on a target recognition type set on the CT image feature set and the MRI image feature set, respectively, to obtain a first confidence sequence and a second confidence sequence.

[0114] Furthermore, in an embodiment of the present invention, if the confidence level of the first confidence level sequence for the target recognition type of neural tissue is 20%, and the confidence level of the second confidence level sequence for the target recognition type of neural tissue is 90%, then the feature allocation weights for neural tissue are 2:9. After the feature allocation weights for target recognition types such as blood vessels, bones, and soft tissue are determined, a weight configuration set is obtained.

[0115] Specifically, in an embodiment of the present invention, when the image recognition model wants to identify neural tissue in CT scan images and MRI images, it assigns 2 weights to the CT image feature set and 9 weights to the MRI image feature set, thereby performing recognition and obtaining the identified neural tissue. Similarly, after identifying tissues such as bones and muscles, a target recognition result set is obtained. Then, the present invention obtains positional distribution information by examining the positional relationship of each target recognition result in the target recognition result set, and performs a three-dimensional arrangement using a drawing tool to obtain a three-dimensional lesion area structure diagram.

[0116] S2. Performing a feature recognition operation on the three-dimensional lesion region structure image to obtain a set of tissue features surrounding the lesion.

[0117] The feature recognition operation is the same as the feature extraction operation for CT scan images and MRI images in S1, except that the processing object here is the three-dimensional lesion region structure map. The set of perilesion tissue features is the feature extraction result of the three-dimensional lesion region structure map.

[0118] Specifically, in an embodiment of the present invention, performing a feature recognition operation on the three-dimensional lesion region structure image to obtain a set of lesion surrounding tissue features includes:

[0119] Using a pre-constructed convolution kernel set, a traversal convolution operation is performed on the three-dimensional lesion area structure map to obtain a convolution matrix set;

[0120] Performing an average pooling operation on the convolution matrix set to obtain a pooling matrix set;

[0121] A flattening operation is performed on each pooling matrix in the pooling matrix set to obtain a set of perilesion tissue features.

[0122] The convolution kernels in the convolution kernel set are core tools in image processing and deep learning, primarily used to extract features from data. Each convolution kernel extracts a corresponding feature. The traversal convolution operation refers to the traversal execution of each convolution kernel. The convolution matrix set is the feature extraction result of the convolution kernel set.

[0123] The average pooling operation is an important downsampling operation in convolutional neural networks. Its core function is to compress data dimensions, preserve overall features, and suppress noise by calculating the mean of local regions. The pooling matrix set refers to the average pooling result of the convolution matrix set.

[0124] The flattening operation refers to the process of further splitting and splicing the pooling matrix to reduce its dimension to 1 dimension.

[0125] Specifically, in an embodiment of the present invention, feature extraction is first performed on the three-dimensional lesion region structure map using a 3×3 or 5×5 convolution kernel specifically designed to extract features from the target recognition result set. The present invention converts the type and distribution information of tissue in each region into feature vectors and performs convolution to obtain a set of convolution matrices. Then, two dimensionality reduction operations, average pooling and flattening, are performed sequentially to obtain a set of pooling matrices and a set of perilesional tissue features.

[0126] S3. Perform a minimally invasive surgery plan identification operation on the tissue feature set surrounding the lesion to obtain a predicted channel number.

[0127] The minimally invasive surgery plan identification operation refers to the process of identifying the number of channels required to treat the lesion based on the standardized surgical process, lesion distribution, and the complexity of the surrounding environment. Each channel has a different purpose, such as controlling a manipulator, cutting a manipulator, or observing a manipulator. The predicted number of channels refers to the minimally invasive surgery plan identification result based on the set of tissue features surrounding the lesion.

[0128] Specifically, in an embodiment of the present invention, performing a minimally invasive surgery plan identification operation on the set of tissue features surrounding the lesion to obtain a predicted number of channels includes:

[0129] Performing lesion type identification on the set of tissue features surrounding the lesion to obtain the lesion type;

[0130] Using a pre-built surgical standardized channel number table, query the channel number corresponding to the lesion type to obtain an initial channel number;

[0131] Performing lesion complexity identification on the set of tissue features surrounding the lesion to obtain a complexity score;

[0132] According to the complexity score, the initial number of channels is adaptively adjusted to obtain a predicted number of channels.

[0133] The lesion type recognition refers to the process of identifying the lesion type through a neural network. The lesion type is the name of a specific lesion, for example, laparoscopic cholecystectomy or thoracoscopic lung nodule resection.

[0134] The standardized channel number table for surgery refers to the standardized channel number for conventional surgery, and the initial channel number refers to the channel number determined based on the lesion type.

[0135] Lesion complexity identification refers to the process of assessing the difficulty of lesion treatment based on the lesion type, distribution range, and risk level of surrounding tissues using a neural network. The complexity score refers to the result of lesion complexity identification. Adaptive adjustment refers to the process of adaptively adding one or two channels based on the complexity score. The predicted number of channels refers to the number of channels required for puncture determined before surgery.

[0136] Specifically, in the embodiment of the present invention, identifying the lesion type is a simple classification judgment network that can be directly obtained and used to obtain the lesion type.

[0137] Specifically, in an embodiment of the present invention, according to the table of standardized surgical channel numbers, if the lesion type is laparoscopic cholecystectomy, 3-4 channels (umbilical observation port, right upper abdominal main operation port, and subxiphoid auxiliary port) usually need to be opened, and if the lesion type is thoracoscopic lung nodule resection, generally 1 observation port (camera channel) and 1-2 operation ports are established. If complex operations (such as lymph node dissection) are required, the number of ports may be increased to 3.

[0138] Specifically, in an embodiment of the present invention, pre-constructed training samples are used to enable the neural network to learn the mapping relationship between [lesion image] and [actual number of treatment channels]. The mapping relationship between [complexity score] and [actual number of treatment channels] is constructed through the output function in the output layer of the neural network. The neural network can then identify the complexity of the lesion on the set of tissue features surrounding the lesion, obtain a complexity score, and then obtain the predicted number of channels.

[0139] S4. Perform a minimally invasive surgical path identification operation based on the predicted number of channels on the tissue feature set surrounding the lesion to obtain a minimally invasive path set, and obtain necessary tissue avoidance sequences within a preset range area on each minimally invasive path in the minimally invasive path set to obtain a necessary tissue avoidance sequence set.

[0140] The minimally invasive surgical path identification operation refers to the process of automatically generating a short, low-risk minimally invasive path through a neural network. The minimally invasive path set refers to each identified minimally invasive path, and the specific number of its elements is equal to the number of predicted channels.

[0141] The preset range area is configured to be within a range of 2-3 cm around the minimally invasive path.

[0142] The necessary tissue avoidance sequence refers to a tissue sequence that experts and technicians believe should be avoided, such as nerves, bones, and other tissues that are difficult to repair or puncture. The necessary tissue avoidance sequence set refers to a set of necessary tissue avoidance sequences within the preset range.

[0143] In detail, in an embodiment of the present invention, performing a minimally invasive surgical path identification operation based on the predicted number of channels on the set of tissue features surrounding the lesion to obtain a minimally invasive path set includes:

[0144] Obtaining the location of the lesion based on the set of tissue features surrounding the lesion, and obtaining the shortest penetration route between a preset external region and the lesion location;

[0145] Obtaining target recognition results of the shortest penetration route within a preset wound range from the target recognition result set to obtain a damaged target set;

[0146] Using a pre-built avoidance weight table, querying the avoidance score of each damaged target in the damaged target set to obtain an avoidance score set, and obtaining a total avoidance score of the avoidance score set;

[0147] Determining whether the total avoidance score is greater than a preset cutting threshold;

[0148] When the total avoidance score is greater than or equal to the cutting threshold, the shortest penetration path is determined to be unqualified, and the shortest penetration path is fine-tuned according to the pre-built genetic algorithm and the damage target set to obtain an updated penetration path. The updated penetration path is used to replace the shortest penetration path, and the process returns to the above step of obtaining the target recognition result of the shortest penetration path within the preset wound range from the target recognition result set;

[0149] When the total avoidance score is less than the cutting threshold, the shortest penetration path is determined to be qualified, and the shortest penetration path is sent to a pre-built candidate path database to obtain a qualified minimally invasive path;

[0150] A minimally invasive pathway set is obtained from the candidate pathway database.

[0151] The lesion location refers to the central location of the lesion distribution. The extracorporeal area refers to an area within the body where medical equipment can be performed, such as the belly, breast, and other areas. When necessary, non-invasive areas such as the nasal cavity and uterus that are in contact with external equipment can also serve as extracorporeal areas. The shortest penetration route refers to the shortest direction vector from the inside of the body to the outside of the body. For example, if the gallbladder is very close to the right upper abdomen of the belly, puncture can be performed directly from the right upper abdomen to reach the gallbladder directly.

[0152] The wound range is within 1-2 cm around the minimally invasive path. The damaged target set refers to some target recognition results that will be lost when puncturing according to the shortest penetration route.

[0153] The avoidance weight table refers to a table of weights assigned by technicians to tissue structures that should be avoided as much as possible. The avoidance score is the score assigned to each tissue marker in the weight distribution table, for example, 1 for bone, 0.9 for neural tissue, 0.8 for blood vessels larger than 3 mm in diameter, etc. The avoidance score set refers to the set of avoidance scores corresponding to the set of damage targets. The total avoidance score is the integral of the avoidance scores corresponding to the penetration length of each tissue on the shortest penetration path.

[0154] The cutting threshold refers to the dividing line that distinguishes whether each position can be punctured.

[0155] Among them, the path fine-tuning means that since the shortest penetration route will pass through various tissues, the shortest penetration route can be segmented according to tissue distribution, such as the first segment, the second segment, and the third segment. If the second segment passes through a blood vessel with a diameter greater than 3 mm, the second segment can be bent to avoid the operation of the blood vessel.

[0156] The genetic algorithm is a global optimization search method that simulates the biological evolution process, searching for the optimal solution in the solution space through a "survival of the fittest" mechanism. Based on the set of avoidance scores, the present invention enables the genetic algorithm to adjust a path toward a smaller total avoidance score, thereby improving the efficiency of the path fine-tuning process. The updated penetration route refers to the shortest penetration route after path fine-tuning.

[0157] The replacement refers to the process of overwriting the content in the original shortest penetration route object with the content in the updated penetration route.

[0158] The candidate pathway database refers to a database storing qualified minimally invasive pathways.

[0159] Specifically, in an embodiment of the present invention, a neural network is used to directly obtain the location of a lesion, such as its relative coordinates relative to a key location on the human body, such as the navel. The shortest penetration path between a predetermined external region and the lesion location is then determined, such as the path from an internal organ to the abdomen. Because the puncture process requires cutting, all tissue within 1-2 cm of the shortest penetration path is severed. The collection of these severed tissues forms the damage target set.

[0160] Although puncture will always damage some tissue cells, some tissue cells are difficult to recover and the risk is relatively high. Therefore, it is necessary to use a pre-built avoidance weight table to analyze the avoidance scores of each damage target in the damage target set, obtain the avoidance score set, and then obtain the total avoidance score.

[0161] When the total avoidance score exceeds the cutting threshold, it indicates that the shortest penetration path is high-risk and unqualified as a minimally invasive path, thus requiring adjustment. The present invention uses a genetic algorithm to avoid the largest one or more damage targets in the damage target set, thereby achieving path fine-tuning and obtaining an updated penetration path. The present invention requires the updated penetration path to be evaluated in the same way as the shortest penetration path. Therefore, by replacing the shortest penetration path with the updated penetration path and then constructing a loop of path adjustment and evaluation, the efficiency of path fine-tuning can be improved.

[0162] When the total avoidance score is less than the cutting threshold, it indicates that the risk of puncture is not high and it is a qualified minimally invasive path. Therefore, it can be sent to the candidate path database to obtain a minimally invasive path set for reference by medical staff.

[0163] Specifically, in an embodiment of the present invention, obtaining a minimally invasive pathway set from the candidate pathway database includes:

[0164] Obtaining minimally invasive paths in the candidate path database to obtain stored paths and the number of paths;

[0165] Determining whether the number of paths is less than the predicted number of channels;

[0166] When the number of paths is less than the predicted number of channels, performing a minimally invasive surgical path identification operation based on the predicted number of channels on the set of tissue features surrounding the lesion, obtaining a minimally invasive path other than the stored path;

[0167] When the number of paths is equal to the number of predicted channels, the stored paths are output to obtain a minimally invasive path set.

[0168] The stored paths refer to the minimally invasive paths in the candidate path database, and the number of paths refers to the number of stored paths.

[0169] Specifically, in this embodiment of the present invention, since a single lesion may require multiple minimally invasive pathways, each with its own control, cutting, and observation devices, the search for additional minimally invasive pathways beyond the previously identified pathway is continued. Ultimately, when the number of pathways equals the predicted number of channels, the search for additional minimally invasive pathways ceases, resulting in a minimally invasive pathway set.

[0170] Furthermore, in an embodiment of the present invention, it is necessary to obtain a necessary avoidance tissue sequence within 2-3 cm on each minimally invasive path. These necessary avoidance tissues are tissues that are not within the cutting path range, but may be damaged due to errors during the cutting process. Therefore, the present invention needs to evaluate the surrounding environment of each minimally invasive path in the necessary avoidance tissue sequence set.

[0171] S5. Perform a path usage allocation identification operation on each minimally invasive path in the minimally invasive path set according to the necessary avoidance tissue sequence set to obtain a path allocation usage type set.

[0172] The path usage allocation identification operation refers to the process of allocating different treatment instruments to each minimally invasive path. The path allocation usage type set refers to the path usage allocation identification results of each minimally invasive path.

[0173] Specifically, in an embodiment of the present invention, a neural network model can be used to sort the puncture ranges of various instruments, and instruments with the smallest penetration range, such as those with a cutting range of less than 1 cm, can be placed on the minimally invasive path with the most necessary tissue avoidance sequences, while main operating equipment with a large cutting range of 3-4 cm can be placed on the minimally invasive path with the least necessary tissue avoidance sequences, thereby realizing a path usage allocation and identification operation for each minimally invasive path in the minimally invasive path set, and obtaining a path allocation usage type set.

[0174] S6. Perform electromagnetic positioning on the pre-built cutting and puncturing instrument to obtain real-time instrument position information, assign a usage type set according to the path, obtain the wound range of each minimally invasive path in the minimally invasive path set, and obtain wound range information.

[0175] The cutting and puncturing instrument refers to a medical device used to form a minimally invasive wound.

[0176] The electromagnetic positioning refers to the process of tracking the three-dimensional position of the cutting and puncturing instrument through a high-precision magnetic field. The real-time instrument position information is the real-time position information of the cutting and puncturing instrument in the human body.

[0177] The incision range refers to the size of the actual incision during minimally invasive surgery. The incision range information is a collection of incision ranges on each minimally invasive path.

[0178] Specifically, in this embodiment of the present invention, the purpose of each incision is first determined based on a set of path-assigned usage types, and then incision is performed. However, before incision is performed, it should be understood that since each path has a different purpose, different incision ranges are required. Therefore, it is necessary to collect the incision ranges of each minimally invasive path to obtain incision range information.

[0179] S7. Using the cutting and puncturing instrument, according to the real-time instrument position information, the minimally invasive path set, and the wound range information, perform a cutting operation on the target treatment area to obtain a minimally invasive channel set.

[0180] The cutting operation is the process of guiding the cutting robot to perform work. The minimally invasive channel set refers to the set of gaps left on the minimally invasive path after the cutting robot performs the operation.

[0181] Specifically, in an embodiment of the present invention, the inspiration position of the cutting operation is first found according to the minimally invasive path set, and then the cutting and puncturing instrument is used to cut and puncture along the minimally invasive path according to the real-time mechanical position information, and finally reaches the lesion, leaving behind a minimally invasive channel set, thereby completing the navigation work of the minimally invasive surgery.

[0182] In detail, in an embodiment of the present invention, after obtaining the minimally invasive channel set, the method further includes:

[0183] Acquire a pre-built surgical instrument set, perform electromagnetic positioning on the surgical instrument set, and obtain a surgical instrument position set;

[0184] Obtaining the instrument type of each surgical instrument in the surgical instrument set to obtain an instrument type set;

[0185] According to the instrument type set, the minimally invasive path set and the surgical instrument position set, each surgical instrument in the surgical instrument set is passed through a corresponding minimally invasive channel in the minimally invasive channel set.

[0186] The surgical instrument set refers to instruments used to directly treat lesions.

[0187] The electromagnetic positioning is the same as the positioning method for the cutting and puncturing instrument.

[0188] The instrument type refers to the type of each surgical instrument, such as an ultrasonic scalpel, a control arm, an observation device, etc.

[0189] Wherein, passing through the corresponding minimally invasive channel in the minimally invasive channel set means moving the surgical instrument along the direction of the pore lines in the minimally invasive channel without opening a new wound.

[0190] Specifically, in the embodiment of the present invention, after the minimally invasive channel is constructed, various medical devices can be introduced. However, during the introduction of surgical instruments, the correspondence between each surgical instrument and the minimally invasive channel must be distinguished.

[0191] The present invention solves the problems described in the background technology. The present invention uses multi-source data from CT scans and MRI scans to identify the structures surrounding the lesions in the target treatment area and obtain a three-dimensional lesion area structure map. Among them, CT scans focus more on the identification of bones and blood vessel distribution, while MRI scans focus more on the identification of lesions, nerves, and soft tissues. Through the collaborative identification of the two, the clarity and accuracy of the three-dimensional lesion area structure map are greatly improved. Then, based on the set of tissue features surrounding the lesions in the three-dimensional lesion area structure map, the present invention identifies multiple channels at a time, and then, based on the distribution of different tissue features involved in each minimally invasive path, assigns different minimally invasive paths to the task requirements of different medical devices, obtains a minimally invasive path set and a path allocation use type set, thereby effectively providing a route for the subsequent puncture process of the cutting and puncturing instrument, and then guiding the cutting and puncturing instrument to puncture the lesion area. Therefore, the present invention can improve the accuracy of minimally invasive surgical navigation and improve the safety of minimally invasive surgery.

[0192] like Figure 2 , which is a functional module diagram of a gynecological surgery navigation system based on minimally invasive technology provided by one embodiment of the present invention.

[0193] The minimally invasive gynecological surgical navigation system 100 described in the present invention can be installed in an electronic device. Depending on the functionality implemented, the minimally invasive gynecological surgical navigation system 100 can include an information acquisition module 101, a minimally invasive path positioning and allocation module 102, and a minimally invasive channel guidance and puncture module 103. A module, also referred to as a unit, is a series of computer program segments that can be executed by an electronic device processor and perform a fixed function, and is stored in the electronic device's memory.

[0194] The information acquisition module 101 is used to obtain a CT scan image and an MRI image of a preset target treatment area, perform three-dimensional modeling on the CT scan image and the MRI image to obtain a three-dimensional lesion area structure diagram, and perform feature recognition operations on the three-dimensional lesion area structure diagram to obtain a set of tissue features surrounding the lesion;

[0195] The minimally invasive path positioning and allocation module 102 is configured to perform a minimally invasive surgical plan identification operation on the set of tissue features surrounding the lesion to obtain a predicted number of channels, perform a minimally invasive surgical path identification operation based on the predicted number of channels on the set of tissue features surrounding the lesion to obtain a minimally invasive path set, obtain a necessary tissue avoidance sequence within a preset range area on each minimally invasive path in the minimally invasive path set to obtain a necessary tissue avoidance sequence set, and perform a path usage allocation identification operation on each minimally invasive path in the minimally invasive path set based on the necessary tissue avoidance sequence set to obtain a path allocation usage type set;

[0196] The minimally invasive channel guiding puncture module 103 is used to perform electromagnetic positioning on the pre-constructed cutting and puncturing instrument to obtain real-time instrument position information, allocate a usage type set according to the path, obtain the wound range of each minimally invasive path in the minimally invasive path set, obtain wound range information, and use the cutting and puncturing instrument to perform a cutting operation on the target treatment area according to the real-time instrument position information, the minimally invasive path set and the wound range information to obtain a minimally invasive channel set.

[0197] In detail, the modules in the minimally invasive gynecological surgery navigation system 100 according to the embodiment of the present invention are used in the same manner as above. Figure 1 The same technical means are used as the gynecological surgical navigation method based on minimally invasive technology described in , and can produce the same technical effects, so they will not be repeated here.

[0198] like Figure 3 FIG. 1 is a schematic diagram of the structure of an electronic device for implementing a gynecological surgery navigation method based on minimally invasive technology provided by an embodiment of the present invention.

[0199] The electronic device 1 may include a processor 10, a memory 11 and a bus 12, and may also include a computer program stored in the memory 11 and executable on the processor 10, such as a gynecological surgery navigation method program based on minimally invasive technology.

[0200] The memory 11 includes at least one type of readable storage medium, and the readable storage medium includes a flash memory, a mobile hard disk, a multimedia card, a card-type memory (for example, an SD or DX memory, etc.), a magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of the electronic device 1, such as a mobile hard disk of the electronic device 1. In other embodiments, the memory 11 can also be an external storage device of the electronic device 1, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device 1. Furthermore, the memory 11 also includes an internal storage unit of the electronic device 1 and an external storage device. The memory 11 can not only be used to store application software and various types of data installed on the electronic device 1, such as the code of a gynecological surgery navigation method program based on minimally invasive technology, but can also be used to temporarily store data that has been output or is to be output.

[0201] In some embodiments, the processor 10 may be composed of an integrated circuit, such as a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and a combination of various control chips. The processor 10 is the control core (Control Unit) of the electronic device, connecting the various components of the entire electronic device using various interfaces and circuits. It executes or runs programs or modules stored in the memory 11 (such as a gynecological surgery navigation method program based on minimally invasive technology) and calls data stored in the memory 11 to perform various functions of the electronic device 1 and process data.

[0202] The bus 12 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus 12 may be divided into an address bus, a data bus, a control bus, etc. The bus 12 is configured to enable communication between the memory 11 and at least one processor 10, etc.

[0203] Figure 3 Only the electronic device with components is shown, and it can be understood by those skilled in the art that Figure 3 The structure shown does not constitute a limitation on the electronic device 1 , and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0204] For example, although not shown, the electronic device 1 may further include a power source (such as a battery) for powering the various components. Preferably, the power source may be logically connected to the at least one processor 10 via a power management device, thereby implementing functions such as charging management, discharging management, and power consumption management through the power management device. The power source may further include any components such as one or more DC or AC power sources, a recharging device, a power failure detection circuit, a power converter or inverter, a power status indicator, etc. The electronic device 1 may further include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0205] Furthermore, the electronic device 1 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the electronic device 1 and other electronic devices.

[0206] Optionally, the electronic device 1 may further include a user interface, which may be a display or an input unit (such as a keyboard). Optionally, the user interface may also be a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touch device. The display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the electronic device 1 and to display a visual user interface.

[0207] The gynecological surgery navigation method program based on minimally invasive technology stored in the memory 11 of the electronic device 1 is a combination of multiple instructions. When running in the processor 10, it can achieve the following:

[0208] Obtaining a CT scan image and an MRI image of a preset target treatment area, performing three-dimensional modeling on the CT scan image and the MRI image to obtain a three-dimensional lesion area structure diagram;

[0209] Performing a feature recognition operation on the three-dimensional lesion region structure image to obtain a set of tissue features surrounding the lesion;

[0210] Performing a minimally invasive surgery plan recognition operation on the set of tissue features surrounding the lesion to obtain a predicted channel number;

[0211] performing a minimally invasive surgical path identification operation based on the predicted number of channels on the set of tissue features surrounding the lesion to obtain a minimally invasive path set, and obtaining a necessary tissue avoidance sequence within a preset range area on each minimally invasive path in the minimally invasive path set to obtain a necessary tissue avoidance sequence set;

[0212] performing a path usage allocation identification operation on each minimally invasive path in the minimally invasive path set according to the necessary avoidance tissue sequence set to obtain a path allocation usage type set;

[0213] Perform electromagnetic positioning on the pre-built cutting and puncturing instrument to obtain real-time instrument position information, assign a usage type set according to the path, obtain the wound range of each minimally invasive path in the minimally invasive path set, and obtain wound range information;

[0214] The cutting and puncturing instrument is used to perform a cutting operation on the target treatment area according to the real-time instrument position information, the minimally invasive path set and the wound range information to obtain a minimally invasive channel set.

[0215] Specifically, the specific implementation method of the processor 10 for the above instructions can refer to Figures 1 to 3The description of the relevant steps in the corresponding embodiments will not be repeated here.

[0216] Furthermore, if the modules / units integrated in the electronic device 1 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard drive, a magnetic disk, an optical disk, a computer memory, or a read-only memory (ROM).

[0217] The present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program. When the computer program is executed by a processor of an electronic device, the computer program can implement:

[0218] Obtaining a CT scan image and an MRI image of a preset target treatment area, performing three-dimensional modeling on the CT scan image and the MRI image to obtain a three-dimensional lesion area structure diagram;

[0219] Performing a feature recognition operation on the three-dimensional lesion region structure image to obtain a set of tissue features surrounding the lesion;

[0220] Performing a minimally invasive surgery plan recognition operation on the set of tissue features surrounding the lesion to obtain a predicted channel number;

[0221] performing a minimally invasive surgical path identification operation based on the predicted number of channels on the set of tissue features surrounding the lesion to obtain a minimally invasive path set, and obtaining a necessary tissue avoidance sequence within a preset range area on each minimally invasive path in the minimally invasive path set to obtain a necessary tissue avoidance sequence set;

[0222] performing a path usage allocation identification operation on each minimally invasive path in the minimally invasive path set according to the necessary avoidance tissue sequence set to obtain a path allocation usage type set;

[0223] Perform electromagnetic positioning on the pre-built cutting and puncturing instrument to obtain real-time instrument position information, assign a usage type set according to the path, obtain the wound range of each minimally invasive path in the minimally invasive path set, and obtain wound range information;

[0224] The cutting and puncturing instrument is used to perform a cutting operation on the target treatment area according to the real-time instrument position information, the minimally invasive path set and the wound range information to obtain a minimally invasive channel set.

[0225] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, systems and methods can be implemented in other ways. For example, the system embodiments described above are only exemplary, and actual implementations may have other division methods.

[0226] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed across multiple network elements. Some or all of the modules may be selected to achieve the purpose of the solution of this embodiment according to actual needs.

[0227] In addition, the functional modules in various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or hardware plus software functional modules.

[0228] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0229] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A gynecological surgery navigation method based on minimally invasive technology, characterized in that: The method comprises: Obtaining a CT scan image and an MRI image of a preset target treatment area, performing three-dimensional modeling on the CT scan image and the MRI image to obtain a three-dimensional lesion area structure diagram; Performing a feature recognition operation on the three-dimensional lesion region structure image to obtain a set of tissue features surrounding the lesion; performing a minimally invasive surgery plan recognition operation on the set of tissue features surrounding the lesion to obtain a predicted channel number; performing a minimally invasive surgical path identification operation based on the predicted number of channels on the set of tissue features surrounding the lesion to obtain a minimally invasive path set, and obtaining a necessary tissue avoidance sequence within a preset range area on each minimally invasive path in the minimally invasive path set to obtain a necessary tissue avoidance sequence set; performing a path usage allocation identification operation on each minimally invasive path in the minimally invasive path set according to the necessary avoidance tissue sequence set to obtain a path allocation usage type set; Perform electromagnetic positioning on the pre-built cutting and puncturing instrument to obtain real-time instrument position information, assign a usage type set according to the path, obtain the wound range of each minimally invasive path in the minimally invasive path set, and obtain wound range information; The cutting and puncturing instrument is used to perform a cutting operation on the target treatment area according to the real-time instrument position information, the minimally invasive path set and the wound range information to obtain a minimally invasive channel set.

2. The gynecological surgery navigation method based on minimally invasive technology according to claim 1, characterized in that: The step of obtaining a CT scan image and an MRI image of a preset target treatment area includes: Using a pre-built CT scanning device, a CT scan is performed on a preset target treatment area to obtain an initial CT scan image, and using a pre-built nuclear magnetic resonance device, a nuclear magnetic resonance scan is performed on the target treatment area to obtain an initial MRI image; performing Gaussian filtering on the initial CT scan image to obtain a noise-reduced CT image, performing frequency domain processing based on Fourier transform on the noise-reduced CT image to obtain an artifact-corrected CT image, and performing CT value normalization on the artifact-corrected CT image to obtain a standard CT image; Performing anisotropic denoising on the initial MRI image to obtain a denoised MRI image, and performing spatial normalization on the denoised MRI image according to a preset standard template to obtain a standard MRI image; The standard CT image and the standard MRI image are aligned using a pre-built mutual information algorithm to obtain a CT scan image and an MRI image.

3. The gynecological surgery navigation method based on minimally invasive technology according to claim 2, characterized in that: The three-dimensional modeling of the CT scan image and the MRI image to obtain a three-dimensional lesion area structure diagram includes: Using a pre-built image recognition model, performing a feature extraction operation on the CT scan image to obtain a CT image feature set, and performing a feature extraction operation on the MRI image to obtain an MRI image feature set; performing feature importance analysis on the CT image feature set according to a preset target recognition type set to obtain a first confidence sequence; Performing feature importance analysis on the MRI image feature set to obtain a second confidence sequence; Obtaining, based on the first confidence sequence and the second confidence sequence, a feature allocation weight for each target recognition type in the target recognition type set to obtain a weight configuration set; performing a fully connected recognition operation based on the target recognition type set on the CT image feature set and the MRI image feature set according to the weight configuration set to obtain a target recognition result set; The positional relationship of each target recognition result in the target recognition result set is obtained to obtain position distribution information, and the target recognition result set is arranged in three dimensions according to the position distribution information to obtain a three-dimensional lesion area structure map.

4. The gynecological surgery navigation method based on minimally invasive technology according to claim 3, characterized in that: The performing of a feature recognition operation on the three-dimensional lesion region structure image to obtain a set of lesion surrounding tissue features includes: Using a pre-constructed convolution kernel set, a traversal convolution operation is performed on the three-dimensional lesion area structure map to obtain a convolution matrix set; Performing an average pooling operation on the convolution matrix set to obtain a pooling matrix set; A flattening operation is performed on each pooling matrix in the pooling matrix set to obtain a set of perilesion tissue features.

5. The gynecological surgery navigation method based on minimally invasive technology according to claim 4, characterized in that: The performing of a minimally invasive surgery plan identification operation on the tissue feature set surrounding the lesion to obtain a predicted number of channels includes: Performing lesion type identification on the set of tissue features surrounding the lesion to obtain the lesion type; Using a pre-built surgical standardized channel number table, query the channel number corresponding to the lesion type to obtain an initial channel number; Performing lesion complexity identification on the set of tissue features surrounding the lesion to obtain a complexity score; According to the complexity score, the initial number of channels is adaptively adjusted to obtain a predicted number of channels.

6. The gynecological surgery navigation method based on minimally invasive technology according to claim 5, characterized in that: The performing of a minimally invasive surgical path identification operation based on the predicted number of channels on the tissue feature set surrounding the lesion to obtain a minimally invasive path set includes: Obtaining the location of the lesion based on the set of tissue features surrounding the lesion, and obtaining the shortest penetration route between a preset external region and the lesion location; Obtaining target recognition results of the shortest penetration route within a preset wound range from the target recognition result set to obtain a damaged target set; Using a pre-built avoidance weight table, querying the avoidance score of each damaged target in the damaged target set to obtain an avoidance score set, and obtaining a total avoidance score of the avoidance score set; Determining whether the total avoidance score is greater than a preset cutting threshold; When the total avoidance score is greater than or equal to the cutting threshold, the shortest penetration path is determined to be unqualified, and the shortest penetration path is fine-tuned according to the pre-built genetic algorithm and the damage target set to obtain an updated penetration path. The updated penetration path is used to replace the shortest penetration path, and the process returns to the above step of obtaining a target recognition result of the shortest penetration path within the preset wound range from the target recognition result set; When the total avoidance score is less than the cutting threshold, the shortest penetration path is determined to be qualified, and the shortest penetration path is sent to a pre-built candidate path database to obtain a qualified minimally invasive path; A minimally invasive pathway set is obtained from the candidate pathway database.

7. The gynecological surgery navigation method based on minimally invasive technology according to claim 6, characterized in that: The step of obtaining a minimally invasive pathway set from the candidate pathway database includes: Obtaining minimally invasive paths in the candidate path database to obtain stored paths and the number of paths; Determining whether the number of paths is less than the predicted number of channels; When the number of paths is less than the predicted number of channels, performing a minimally invasive surgical path identification operation based on the predicted number of channels on the set of tissue features surrounding the lesion, obtaining a minimally invasive path other than the stored path; When the number of paths is equal to the number of predicted channels, the stored paths are output to obtain a minimally invasive path set.

8. The gynecological surgery navigation method based on minimally invasive technology according to claim 7, characterized in that: After obtaining the minimally invasive channel set, the method further includes: Acquire a pre-built surgical instrument set, perform electromagnetic positioning on the surgical instrument set, and obtain a surgical instrument position set; Obtaining the instrument type of each surgical instrument in the surgical instrument set to obtain an instrument type set; According to the instrument type set, the minimally invasive path set and the surgical instrument position set, each surgical instrument in the surgical instrument set is passed through a corresponding minimally invasive channel in the minimally invasive channel set.

9. A gynecological surgery navigation system based on minimally invasive technology, characterized in that: The system comprises: An information acquisition module is configured to acquire a CT scan image and an MRI image of a preset target treatment area, perform three-dimensional modeling on the CT scan image and the MRI image to obtain a three-dimensional lesion area structure diagram, and perform feature recognition on the three-dimensional lesion area structure diagram to obtain a set of tissue features surrounding the lesion; a minimally invasive path positioning and allocation module, configured to perform a minimally invasive surgical plan identification operation on the set of tissue features surrounding the lesion to obtain a predicted number of channels, perform a minimally invasive surgical path identification operation based on the predicted number of channels on the set of tissue features surrounding the lesion to obtain a minimally invasive path set, obtain a necessary tissue avoidance sequence within a preset range area on each minimally invasive path in the minimally invasive path set to obtain a necessary tissue avoidance sequence set, and perform a path usage allocation identification operation on each minimally invasive path in the minimally invasive path set based on the necessary tissue avoidance sequence set to obtain a path allocation usage type set; The minimally invasive channel guided puncture module is used to electromagnetically locate the pre-built cutting and puncture instrument to obtain real-time instrument position information, allocate a use type set according to the path, obtain the wound range of each minimally invasive path in the minimally invasive path set, obtain wound range information, and use the cutting and puncture instrument to perform a cutting operation on the target treatment area according to the real-time instrument position information, the minimally invasive path set and the wound range information to obtain a minimally invasive channel set.

10. The gynecological surgery navigation system based on minimally invasive technology according to claim 9, characterized in that: The performing of a minimally invasive surgical path identification operation based on the predicted number of channels on the tissue feature set surrounding the lesion to obtain a minimally invasive path set includes: Obtaining the location of the lesion based on the set of tissue features surrounding the lesion, and obtaining the shortest penetration route between a preset external region and the lesion location; Obtaining target recognition results of the shortest penetration route within a preset wound range from the target recognition result set to obtain a damaged target set; Using a pre-built avoidance weight table, querying the avoidance score of each damaged target in the damaged target set to obtain an avoidance score set, and obtaining a total avoidance score of the avoidance score set; Determining whether the total avoidance score is greater than a preset cutting threshold; When the total avoidance score is greater than or equal to the cutting threshold, the shortest penetration path is determined to be unqualified, and the shortest penetration path is fine-tuned according to the pre-built genetic algorithm and the damage target set to obtain an updated penetration path. The updated penetration path is used to replace the shortest penetration path, and the process returns to the above step of obtaining a target recognition result of the shortest penetration path within the preset wound range from the target recognition result set; When the total avoidance score is less than the cutting threshold, the shortest penetration path is determined to be qualified, and the shortest penetration path is sent to a pre-built candidate path database to obtain a qualified minimally invasive path; A minimally invasive pathway set is obtained from the candidate pathway database.

Images