Gynecological pelvic and abdominal tumor operation training result discrimination method and anatomical model

Abstract

The invention relates to a gynecological pelvic and abdominal tumor operation training result discrimination method and an anatomical model. The method comprises the following steps: acquiring video data and a response signal list in an operation training process; the response signal list comprises response signal streams of a plurality of structure nodes of the anatomical model; extracting data features of the video data to obtain operation features; the operation characteristics comprise a structure tracking matrix and an action track sequence; performing time sequence processing on the response signal flow of the response signal list to obtain a structure response state table; the structure response state table comprises a structure number, a response state and response time; generating an operation process judgment report according to the operation characteristics and the structure response state table; and based on a preset score table, performing scoring according to the operation process judgment report to obtain an operation training result score. By adopting the method, the operation training completion degree and quality can be automatically output without relying on subjective judgment.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of medical training, and particularly relates to a gynecological pelvic-abdominal tumor surgery training result discrimination method and an anatomical model. BACKGROUND

[0002] With the development of medical education and surgical training technology, various surgical training methods are introduced in medical teaching to improve the cognitive ability of scholars on complex anatomical regions and the level of surgical skill. In the treatment of gynecological malignant tumors, extensive hysterectomy, pelvic lymph node dissection, abdominal aortic lymph node dissection and other high-risk and high-complexity operations are involved, which puts forward higher requirements for the ability of the surgeon to identify anatomical structures and plan surgical paths.

[0003] In the traditional technology, the training effect depends on the artificial observation and scoring of teachers, which is usually evaluated through on-site inquiry, naked eye inspection or video playback after the training. Although this method can provide certain guidance value, the evaluation dimension is limited, the result lacks objective standard, and the evaluation granularity is relatively coarse, which is difficult to reflect the accuracy of detailed operation and the precision of structure identification. The trainees are difficult to obtain guidance on detailed operation, especially in terms of cleaning range and operation path standardization, and there is no verifiable basis for judgment. Whether the important tissues involved in the training are accurately identified and processed cannot be accurately judged. SUMMARY

[0004] Therefore, it is necessary to provide a gynecological pelvic-abdominal tumor surgery training result discrimination method and an anatomical model which can objectively discriminate the surgical training effect in view of the above technical problems.

[0005] In a first aspect, the present application provides a gynecological pelvic-abdominal tumor surgery training result discrimination method applied to a server in a gynecological pelvic-abdominal tumor surgery training result discrimination system, wherein the gynecological pelvic-abdominal tumor surgery training result discrimination system is connected with an anatomical model and a camera; the anatomical model simulates the pelvic-abdominal cavity structure of the human body and is used to provide a training personnel for pelvic-abdominal tumor surgery training, and the anatomical model is integrated with a response module for collecting a surgery process response signal stream; the response module comprises a sensor assembly and a communication node; the response module is connected with the server through a built-in first data transmission module; the camera is connected with the server through a built-in second transmission module.

[0006] The gynecological pelvic-abdominal tumor surgery training result discrimination method comprises the following steps.

[0007] Obtaining video data and a response signal list in the surgery training process; the response signal list comprises response signal streams of multiple structure nodes of the anatomical model;

[0008] Extracting data features of the video data to obtain operation features; the operation features comprise a structure tracking matrix and a motion trajectory sequence.

[0009] The response signal stream of the response signal list is subjected to timing processing to obtain a structure response state table; the structure response state table comprises a structure number, a response state, and a response time;

[0010] A surgical procedure judgment report is generated according to the operation characteristics and the structure response state table;

[0011] A surgical training result score is obtained by scoring according to the surgical procedure judgment report based on a preset scoring table.

[0012] In one of the embodiments, data characteristics of the video data are extracted to obtain the operation characteristics, which include:

[0013] Key frame extraction is performed on the video data to obtain an anatomical model image sequence and a motion occurrence frame image sequence;

[0014] An initial state and a final state of the tissue structure of the anatomical model image sequence are identified using a color segmentation algorithm to obtain a structure tracking matrix; the structure tracking matrix includes a tissue structure boundary, a corresponding change characteristic, and a timestamp;

[0015] Surgical instrument detection is performed on the motion occurrence frame image sequence based on optical flow analysis to obtain an instrument motion trajectory;

[0016] The instrument motion trajectory and the structure tracking matrix are aligned according to the timestamp, and behavior type identification is performed to obtain a motion trajectory sequence; the motion trajectory sequence includes a kinematic trajectory and a behavior type;

[0017] The result tracking matrix and the motion trajectory sequence are determined as the operation characteristics.

[0018] In one of the embodiments, the surgical procedure judgment report is generated according to the operation characteristics and the structure response state table, which includes:

[0019] The tissue structure contact sequence is reconstructed according to the structure tracking matrix to obtain an operation path discrimination report;

[0020] The motion trajectory sequence and the structure response state table are aligned according to the response time to obtain a false operation marker list;

[0021] A surgical target hit result is generated according to the operation characteristics and the structure response state list based on a target tissue structure cleaning list; the surgical target hit result includes a cleaning object comparison table, a cleaning completion rate, and a misoperation structure table;

[0022] The surgical procedure judgment report is generated according to the operation path discrimination report, the false operation marker list, and the surgical target hit result.

[0023] In one of the embodiments, the structure tracking matrix is used to reconstruct the organization structure contact sequence to obtain an operation path judgment report, including:

[0024] The organization structure boundary corresponding to the change feature is arranged in a time-stamped sequential order to obtain an organization structure change sequence;

[0025] The organization structure contact sequence is reconstructed according to the organization structure change sequence to obtain a surgical procedure operation sequence;

[0026] The surgical procedure operation sequence is determined based on the edit distance algorithm whether it conforms to the surgical procedure template path to obtain an operation path judgment report; the operation path judgment report includes path completeness, path interruption and path error.

[0027] In one of the embodiments, the action trajectory sequence and the structure response state table are aligned according to the response time to obtain a false operation marking list, including:

[0028] The behavior type contained in the action trajectory sequence is subjected to surgical action screening to obtain an organization structure interaction action trajectory;

[0029] The time sequence corresponding to the organization structure interaction action trajectory is aligned with the response time to obtain an alignment result; the alignment result includes a corresponding response state event and a non-corresponding response state event;

[0030] If the alignment result is a non-corresponding response state event, the organization structure interaction action trajectory is marked as a false operation, and a false operation marking list is generated.

[0031] In one of the embodiments, based on the target organization structure cleaning list, the operation feature and the structure response state list are used to generate a surgical target hit result, including:

[0032] The structure tracking matrix, the action trajectory sequence and the response state are causally associated according to a preset time window to obtain a cleaning object table;

[0033] The cleaning object table is compared with the target organization structure cleaning list to obtain a cleaning object comparison table; the cleaning object comparison table includes a target organization structure determination result;

[0034] According to the structure number and the response state, the area corresponding calculation is performed according to the target organization structure cleaning list to obtain the cleaning completion rate of each area;

[0035] The structure response state table and the target organization result cleaning list are used to generate a misoperation structure table.

[0036] In a second aspect, the application also provides a gynecological pelvic cavity tumor surgery training result judgment system, including:

[0037] Anatomy model, server and camera;

[0038] The anatomy model simulates the pelvic-abdominal cavity structure of a human body, and is used to provide training personnel with training of a pelvic-abdominal tumor operation, and the anatomy model is integrated with a response module for collecting a response signal stream of the operation process;

[0039] The response module comprises sensors and a communication node; the response module is connected with the server through a first data transmission module built-in;

[0040] The camera is connected with the server through a second data transmission module built-in;

[0041] The server is used to implement the steps of the gynecological pelvic-abdominal tumor operation training result discrimination method.

[0042] In a third aspect, the application further provides an anatomy model for gynecological pelvic-abdominal tumor operation training, comprising:

[0043] A main body part simulating a pelvic-abdominal cavity structure, an instrument assembly for training widely including hysterectomy, pelvic lymph node dissection and para-aortic lymph node dissection operation, a response module integrated in the main body part and a data transmission unit; the response module comprises a sensor assembly and a communication node;

[0044] The main body part has a simulation tissue module simulating human skin, pelvic tissue structure, abdominal aorta and inferior vena cava, ureter, uterus and its appendages, upper abdominal organs and lymph node groups; the lymph node groups are attached to the pelvic cavity and the abdominal aorta and inferior vena cava; each simulation tissue module of the main body part has different material rigidity to simulate the real human tissue morphology, texture, separability and adhesion degree;

[0045] The sensor assembly comprises pressure sensors for detecting the operation force acting on the simulation tissue module in the resection, traction and cleaning operation, and the pressure sensors are arranged at the lymph node group attachment points, blood vessel walls and anatomical interface positions;

[0046] The communication node comprises a communication node arranged on each anatomical interface to obtain the signals on the anatomical interface and obtain the response signals of each simulation tissue module in the operation training process, for detecting whether the simulation tissue module including the uterus and its appendages, upper abdominal organs and pelvic tissue structure is resected and separated;

[0047] The data transmission unit is connected with the response module, and is used to transmit the data collected by the response module to the server, and the server is used to implement the steps of the gynecological pelvic-abdominal tumor operation training result discrimination method.

[0048] In a fourth aspect, the present application also provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any of the above gynecological pelvic-abdominal tumor surgery training result determination methods when executing the computer program.

[0049] In a fifth aspect, the present application also provides a computer readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any of the above gynecological pelvic-abdominal tumor surgery training result determination methods.

[0050] The gynecological pelvic-abdominal tumor surgery training result determination method and the anatomical model can automatically output the completion degree and quality of the surgery training without relying on subjective judgment by combining structure tracking with response state analysis; can identify "empty operation" or "wrong path" that does not produce tissue response by using multi-modal data alignment and behavior anomaly detection; can quantitatively evaluate the deviation between the operator path and the standard path based on an edit distance algorithm, and provide objective feedback to the trainer; can make targeted judgments on the cleaning effect of key structures such as lymph node groups, and generate a task hit rate index that meets the training purpose; can realize a logical closed loop of action behavior and tissue feedback by fusing data collected by sensors and cameras in the anatomical model, and improve the evaluation accuracy; and the method is suitable for training and evaluation tasks of different gynecological pelvic-abdominal tumor surgery training procedures and anatomical regions, and can be continuously expanded with model upgrades. BRIEF DESCRIPTION OF DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the drawings needed to be used in the embodiments or the related art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0052] Figure 1 The flowchart of the gynecological pelvic-abdominal tumor surgery training result determination method of the present application;

[0053] Figure 2 The flowchart of the sub-step of step S104;

[0054] Figure 3 The composition structure diagram of the gynecological pelvic-abdominal tumor surgery training result determination system of the present application;

[0055] Figure 4 The internal composition structure diagram of the anatomical model for gynecological pelvic-abdominal tumor surgery training;

[0056] Figure 5 The structure diagram of the anatomical model for gynecological pelvic-abdominal tumor surgery training;

[0057] Figure 6 The schematic diagram of the composition of the gynecological pelvic abdominal tumor surgery training dissection model. DETAILED DESCRIPTION

[0058] In order to make the purpose, technical scheme and advantages of the present application more clear, the present application is further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application and do not limit the present application.

[0059] In one embodiment, as shown in Figure 1 A gynecological pelvic abdominal tumor surgery training result discrimination method is provided, which is applied to a server in a gynecological pelvic abdominal tumor surgery training result discrimination system, the gynecological pelvic abdominal tumor surgery training result discrimination system being connected with a dissection model and a camera; the dissection model simulates the pelvic abdominal cavity structure of a human body and is used to provide a training personnel for pelvic abdominal tumor surgery training, and the dissection model is integrated with a response module for collecting a surgery process response signal stream; the response module includes a sensor assembly and a communication node; the response module is connected with the server through a built-in first data transmission module; and the camera is connected with the server through a built-in second transmission module. The method includes the following steps in the embodiment:

[0060] S101, video data and a response signal list in the surgery training process are acquired; the response signal list includes response signal streams of multiple structure nodes of the dissection model.

[0061] Illustratively, the video data refers to continuous image frame sequences collected by external imaging devices such as high-definition cameras, laparoscope simulation lenses and the like matched with the system in the whole surgery training process, which can be an RGB video stream, and the frame rate is between 25 and 60 fps to ensure complete capture of operation details. The collection device can be fixedly installed above the training model structure or at the end of the laparoscope channel to ensure that the video image can cover the complete operation field of view. The video data not only records the physical action of the training personnel using the instrument, but also completely retains information such as tissue deformation caused by the surgery operation on the structure surface or the dissection junction, instrument path, dissection exposure degree and the like.

[0062] Schematic representation: The response signal list is a real-time data set acquired by sensor components and communication nodes in the anatomical training model, including biomechanical or electrical signal changes in response to surgical procedures by various anatomical structures. It is represented by structural node IDs, signal channels, and time series. The structural node ID (ID) assigns each communication node or sensor a unique identifier to a specific anatomical structure, such as the left round ligament of the uterus, the right internal iliac artery, or the right ureter, for locating the specific structure in the response stream. Signal channels include pressure channels, impedance channels, and contact channels. The time series shows the response signal as a function of time, with a sampling frequency consistent with video frames to ensure time synchronization between multimodal data. Signal values ​​record the actual sampled values ​​of the current channel, including pressure values, contact resistance values, current response, and communication signal connectivity. For example, for the pelvic lymphoid tissue numbered "Node_14", which is subjected to improper traction during the time period t = [0s, 10s], its pressure sensor channel may record as: s(t) = [0.02, 0.05, 0.09, 0.14, 0.17, 0.20, 0.15, 0.08, 0.04, 0.01] (Pa).

[0063] Optionally, the two types of data can be time-aligned during acquisition. Using a unified system clock or reference trigger signal method, a timestamp is broadcast to the data bus simultaneously with each captured video frame, allowing video frames and response signals to be matched using either "frame number" or "unified timestamp". For example, at frame 100, when the video captures the puncture device entering the pelvic region, the sensor channel synchronously records the trigger contact signal value of structure Node_09 as "1", achieving data matching and correspondence.

[0064] S102. Extract the data features from the video data to obtain the operational features; the operational features include the structure tracking matrix and the motion trajectory sequence.

[0065] The video data is structurally analyzed, transforming it from an image sequence into quantifiable operational features suitable for analysis, including the trajectory of surgical instruments and the evolution of spatial relationships between target anatomical structures. The Structure Tracking Matrix (STM) is used to describe the interactions between various structures within the anatomical model in the training video. Illustratively, a structure recognition model based on a deep semantic segmentation network is used to annotate each frame of the image at the pixel level, including different tissue modules within the model such as the uterus, uterine arteries and veins, ureters, pelvic lymphatic chains, common iliac artery, intestines, and abdominal aorta.

[0066] By identifying each frame of the image using the above model, the visual state flag of each structure at time t can be obtained. in denotes the i-th structure can be recognized in frame f t ; otherwise, 0. Further, to improve the timing consistency and exclude image errors, a structure-level Kalman filter or multi-frame feature fusion mechanism is introduced to perform time window smoothing processing on the state of structure visibility, to obtain a structure tracking matrix STM, which is a three-dimensional matrix, in the form of:

[0067]

[0068] This matrix can be regarded as a time axis mapping of the anatomical structure being identified, exposed, contacted and operated during the entire surgical training process. Each structure is recorded in the STM from the first time it is explicitly identified to whether it is subsequently exposed in its entirety, whether it is deformed or missing.

[0069] Illustratively, at the same time as structure recognition, the system performs action trajectory extraction in parallel. The action trajectory sequence refers to the motion trajectory of the trainee's operating instrument in the image, representing its operation path. The end point of the operating instrument can be tracked by means of inter-frame difference, optical flow method or instrument end color marker identification, etc. to form a path vector sequence A = {a1, a2, … a T}, each a T is a spatial coordinate at time t. In combination with the positioning of the anatomical structure, a "structure-action matching mapping" can be further established to determine whether the trainee's operation covers all key structure regions and whether a false path is generated in a non-target region, etc.

[0070] Further, the action trajectory sequence reflects the actions of the instrument such as puncture, traction, clamping and cutting through the spatial movement trajectory of the instrument in the time dimension. T is a coordinate point in three-dimensional space:

[0071]

[0072] The action type is identified by the change characteristics of the three-dimensional spatial coordinate points.

[0073] S103, time series processing is performed on the response signal stream of the response signal list to obtain a structure response state table; the structure response state table includes structure number, response state and response time.

[0074] Illustratively, the received response signal list is: R = {r1(t), r2(t), … r n (t)}, where r i (t) represents the sensing response signal stream of the i-th structure node at time t, which can be single-channel (such as pressure value) or multi-channel joint data. Each structure node can correspond to an explicit tissue structure number in the model, and its physical position matches the structure region in image analysis.

[0075] Specifically, the timing processing includes: pre-cleaning the signal by using sliding window average, wavelet denoising or Butterworth filter. At the same time, a unified normalization strategy is provided for the output data of different sensors,

[0076] The response signal is mapped to a state label. For example, if the interface signal of an organization module is disconnected, it is considered as a cleaning or resection event; if the pressure sensor output of a certain structure node continuously exceeds the threshold value, it is considered as a pressure membrane threshold reaching event. After completing the state determination, the system constructs the response state sequence of each structure number, and forms the final structure response state table, including structure number (ID), first response time, response peak time and final response state.

[0077] Optionally, if the electrical impedance sensor is introduced, the range of tissue resection can also be determined by signal changes. The electrical impedance change rate is calculated by the following formula:

[0078]

[0079] Where Z i (t0) is the initial impedance of the structure, Z i (t end ) is the impedance at the end of training. If ΔZ i >δ, it indicates that the tissue in this area has been substantially resected, which can be used to assist in determining the "cleaning degree".

[0080] S104, generating a surgical procedure determination report according to the operation characteristics and the structure response state table.

[0081] Illustratively, the operation characteristics and the structure response state table are synchronized and aligned. Considering that each frame in the image sequence has a timestamp t, and the response signal stream also has clear time information, the system performs time alignment at the frame level: tracking the structure matrix M(t), where each frame represents the position, area and deformation state of each structure region in the current frame image; the action trajectory sequence A(t) contains the path, speed, dwell time and spatial changes of the surgical instrument passing through the structure in the continuous frames; the structure response state table SRS, through the response time period of each structure, the system can determine whether it corresponds to an operation on a certain trajectory path or a certain structure region, and obtains R i =<ID i ,{M i (t)},{A i (t)},{RS i (t)}>,where ID i represents the structure number, M i (t) is the position change of the structure in the visual layer, A i (t) is the action trajectory of the instrument, and RS i(t) is the physical response state. Further, the system constructs a surgical operation judgment model according to the surgical procedure standard: for example, the "extensive hysterectomy" surgical procedure requires that specific ligaments such as the round ligament of uterus and the cardinal ligament be clearly identified and completely cut; the uterine body tissue should be completely separated, and the response area should contain responses of multiple structure nodes; the cutting process should be completed along the standard trajectory of the instrument, avoiding injury to adjacent tissues such as the ureter; the involved tissue structure should have a cutting or pulling response state, and the duration should meet the conventional experience interval.

[0082] For each surgical procedure step, the system formulates the following judgment indicators:

[0083] Structure recognition rate (SRR): if the structure area in the structure tracking matrix is clearly labeled and has a corresponding response record in the structure response table, it is considered that the structure recognition is complete:

[0084]

[0085] Operation path matching degree (PMD): according to the trajectory deviation between the surgical procedure specified path and the operator's operation path, dynamic time warping (DTW) matching is performed to obtain a matching score:

[0086]

[0087] Tissue response consistency (ORI): whether the structure response overlaps with the operation trajectory, and whether the response state changes in the area passed by the operation trajectory:

[0088]

[0089] Error operation rate (EOR): if the system detects that a structure area that is not required to be processed has a response that is not a "light touch" level in the action trajectory, it is counted as an error operation:

[0090]

[0091] The system constructs an evaluation vector according to the above indicators, generates a scoring table for each surgical procedure step, and generates an operation time axis graph during the surgery, marking the key structure processing time period. The judgment report form can include: structure processing completion table; trajectory matching analysis graph: response consistency heat map: integrity and standardization conclusion: for example, "right internal iliac lymphatic chain" is identified and path covered in the image, but the response state is "no response", then the system will judge it as operation illusion, otherwise it is judged as effective processing.

[0092] S105, based on the preset scoring table, scoring according to the surgical procedure judgment report to obtain the surgical training result score.

[0093] In illustration, the scoring template is a pre-set structured scoring table that sets different scoring indicators for each type of surgical step and assigns weight coefficients. These scoring dimensions may include: Structural Recognition Rate (SRR); Operation Path Matching Degree (PMD); Tissue Response Consistency (ORI); Operation Time Standardization (OPT); Error Rate (EOR); Percentage of Unprocessed Structures (UPS), etc.

[0094] Furthermore, a weighted comprehensive score is calculated according to the following general formula:

[0095]

[0096] Where, ω i represents the weight of the i-th dimension; represents the actual evaluation value of that dimension; represents the scoring function for that dimension, which is usually a linear function or an interval scoring function, as defined by the standard.

[0097]

[0098] For example, for the Operation Path Matching Degree (PMD), the standard range is [0.85, 1.0], the acceptable range is [0.6, 0.85), and a score below 0.6 is recorded as 0.

[0099] The system iterates through and scores all structures, all operational paths, and all response behaviors using the methods described above, then merges these scores to generate a total score. Optionally, the scoring system can be configured as a tiered evaluation mechanism: 90 points and above: sufficient structural cleaning, standardized operation, and high training completion rate; 70-89 points: most structures processed, paths generally reasonable, with minor deviations; 50-69 points: significant structural omissions or path problems, requiring operational optimization; below 50 points: structures not identified or obviously incorrect paths, considered invalid training. If required by users or training institutions, a sub-structure scoring table can also be generated to separately score structural regions such as "hysterectomy," "right internal iliac dissection," and "para-aortic dissection," allowing for more refined feedback.

[0100] In the above gynecological pelvic and abdominal tumor surgery training result determination method, the video data can intuitively reflect the external performance of the operation, such as the movement of surgical instruments, the state of tissues, etc.; and the response signal list records the response signal flow of multiple structure nodes of the anatomical model, which can reflect the influence of the operation on the anatomical structure from the internal mechanism level. This multi-dimensional data acquisition method provides a rich information basis for comprehensive and accurate evaluation of the operation training result. Operation features are extracted from the video data, including a structure tracking matrix and a motion trajectory sequence. The structure tracking matrix can clearly show the position and state change of each anatomical structure during the operation, which helps to analyze the influence of the operation on different structures; the motion trajectory sequence records the movement path of the surgical instrument, making the evaluation of the operation more detailed and in-depth. The operation process determination report is generated according to the operation features and the structure response state table, which can comprehensively consider the external performance of the operation and the internal influence on the anatomical structure, and comprehensively and objectively evaluate the operation training result. The operation process determination report can provide detailed operation evaluation and improvement suggestions for the operator, helping them understand their strengths and weaknesses in the operation training, and make targeted training and improvement. At the same time, the report can also provide a reference for the operation training teachers to adjust the teaching scheme and guide the students' learning.

[0101] In one of the embodiments, data features of the video data are extracted to obtain operation features, including:

[0102] S21, key frame extraction is performed on the video data to obtain an anatomical model image sequence and a motion occurrence frame image sequence.

[0103] Illustratively, the surgery training video often contains a large amount of static or low dynamic information, and redundant frames need to be removed and frames with significant structure changes or motion behaviors need to be retained. Key frame extraction can be screened by frame difference metrics such as inter-frame pixel change rate, structural similarity index SSIM, and histogram difference. Further, the key frames are divided into an anatomical model image sequence and a motion occurrence frame image sequence, wherein the anatomical model image sequence focuses on frames with structure changes of the anatomical model; and the motion occurrence frame image sequence focuses on frames with instruments.

[0104] S22, using a color segmentation algorithm to identify the boundaries corresponding to the initial state and the final state of the tissue structure of the anatomical model image sequence to obtain a structure tracking matrix; the structure tracking matrix includes the tissue structure boundary, the corresponding change feature and the timestamp.

[0105] Illustratively, different colors can be used to distinguish various types of tissue structures during the production of the anatomical model, such as red for blood vessels and yellow for fat, and color segmentation algorithms such as K-means clustering, HSV color space segmentation, region growing method, etc. can be used to achieve segmentation and identification of different tissues. By comparing the images before and after the start of training, the initial boundary and the final boundary of the tissue structure are extracted. In order to quantify the structural changes, the system records the area, position, connectivity changes, and other boundary contour change quantities of each structure, and adds timestamp information, and finally constructs a structure tracking matrix. Each row of the matrix represents a tissue structure unit, and the fields include tissue number, initial boundary, final boundary, change type, change amount indicator and corresponding time period. The change type includes resection, tearing, and no touch.

[0106] S23, based on optical flow analysis, the motion occurs frame image sequence is detected for surgical instruments, and the instrument motion trajectory is obtained.

[0107] Farneback optical flow, PWC-Net depth optical flow and other optical flow algorithms can identify the motion direction and speed of pixels in inter-frame images, thereby obtaining the movement path of the instrument, and the image position of instruments such as surgical forceps and scissors can be located, and the trajectory thereof between multiple frames can be continuously tracked, forming a continuous space-time path, denoted as an instrument motion trajectory, including intra-frame coordinates, timestamps and relative motion speed and other information.

[0108] S24, aligning the instrument motion trajectory with the structure tracking matrix according to the timestamp, and identifying the behavior type to obtain a motion trajectory sequence; the motion trajectory sequence includes a kinematic trajectory and a behavior type.

[0109] The obtained instrument motion trajectory is aligned with the structure tracking matrix according to the time axis. Through timestamp matching, the system can determine the target structure acted on by the instrument in a certain time period. Based on this correlation, the relationship between the motion mode of the instrument, such as continuous linear advancement, cutting action, and rotation pressure, and the corresponding tissue changes is further analyzed, and the behavior type is identified through a rule library. Illustratively, if the instrument maintains a high-speed linear movement in a specific structure area for a short time, and the outline of the structure area is significantly reduced, it can be identified as a "tissue resection" operation; if the trajectory is repeatedly contacted for a short time and the structure outline does not change, it can be judged as a "probing contact" behavior. Finally, each instrument trajectory is assigned a corresponding behavior label to form a motion trajectory sequence, which includes the fields of motion path, behavior type, time period and affected structure.

[0110] S25, determining the result tracking matrix and the motion trajectory sequence as operation features.

[0111] The obtained structure tracking matrix and motion trajectory sequence are merged and stored as operation features.

[0112] In one embodiment, as shown in Figure 2 The surgical procedure decision report is generated based on the operation feature and the structure response state table, including:

[0113] S201, reconstruct the organization structure contact sequence according to the structure tracking matrix, and obtain the operation path discrimination report.

[0114] The system reconstructs the time sequence according to the initial state, final state and timestamp information of each organization structure in the structure tracking matrix, to form the organization structure contact sequence. The contact sequence describes the order of operation or contact of each organization in the surgical training process. By comparing the contact order with the standardized operation path template, that is, the recommended cleaning order or dissection level processing path in the real operation guide, the deviation between the actual path and the standard path can be identified. The deviation includes but is not limited to skipping (missing cleaning), reverse order (disturbing the hierarchical structure) or repetition (low efficiency) and the like. The system summarizes it into an operation path discrimination report, which includes a difference comparison table of the actual contact sequence and the standard contact sequence; the structure number and timestamp of the non-standard operation point; the path deviation score; and the structure damage possibility warning.

[0115] S202, align the action trajectory sequence and the structure response state table according to the response time, to obtain a false operation marking list.

[0116] The system time-aligns the action trajectory sequence and the structure response state table to identify invalid or false operation behaviors. So-called false operation refers to the detection of the operation behavior of the instrument on a certain structure in the action trajectory, but the structure does not show any effective response in the response state table, including no stress change, no change in electrical impedance, or no change in position state. According to the following alignment logic: if there is no corresponding record in the structure response state table within the action time period of the action trajectory, it is marked as "false operation"; if there is a response but the response amplitude is far below the threshold, it is marked as "low-efficiency contact"; if the instrument behavior does not match the structure, such as high-pressure cutting action applied to a non-cleaning target structure, it is marked as "wrong interaction". The final result is output as a false operation marking list, including the behavior number, time period, target structure number, judgment label and possible reason analysis.

[0117] S203, based on the target organization structure cleaning list, generate a surgical target hit result according to the operation feature and the structure response state list; the surgical target hit result includes a cleaning object comparison table, a cleaning completion rate and a misoperation structure table.

[0118] The system evaluates whether the operation hits the target region according to the target tissue structure cleaning list preset by the training task, and forms a surgical target hit result. The list defines the cleaning targets required to be completed in the task, and the system compares these targets with the structure tracking matrix in the operation characteristics, analyzes whether there is a significant change in the structure boundary, and combines the strong response signal in the response state table to confirm whether the operation is effective. For example, if the target structure has "significant deformation + strong response in the corresponding time period" in the tracking matrix, it is determined as "hit the cleaning target"; if there is no operation behavior or weak response, it is marked as "cleaning not completed"; if the operation behavior occurs in a non-target structure and is accompanied by structure deformation and response, it is marked as "misoperation". The final surgical target hit result includes a cleaning object comparison table; cleaning completion rate; and a misoperation structure table.

[0119] S204, generating a surgical process judgment report according to the operation path discrimination report, the false operation mark list and the surgical target hit result.

[0120] In one embodiment, the tissue structure contact sequence is reconstructed according to the structure tracking matrix to obtain the operation path discrimination report, which includes:

[0121] S31, arranging the tissue structure boundary corresponding to the change characteristics in a time-stamp-sequential order to obtain a tissue structure change sequence.

[0122] The system first extracts the boundary change records of all tissue structures from the structure tracking matrix, and arranges them in chronological order according to the time stamp, thereby forming a tissue structure change sequence. Each record in the structure tracking matrix usually contains the following fields: tissue structure number, boundary change characteristics (such as area change, shape distortion, edge displacement, etc.) and time stamp of change. For example, [structure A, boundary shrinkage, T1], [structure B, boundary rupture, T2], [structure C, shape reconstruction, T3] by sorting these change events in the order of T1

[0123] S32, reconstructing a tissue structure contact sequence according to the tissue structure change sequence to obtain a surgical process operation sequence.

[0124] Illustratively, the contact sequence emphasizes the operational logical relationship, i.e., which tissues are contacted at what time point and in what order, different from the change sequence. Specifically, if the boundary of structure B breaks after structure A, but the break of structure B must be preceded by the incision of structure A, then the operation order should be "structure A→structure B". The system will introduce a prior anatomical dependency graph to correct this logic. Illustratively, the original change sequence is: structure C(T1)→structure A(T2)→structure B(T3), but the anatomical dependency graph indicates that structure B must be exposed before structure A, so the operation sequence is corrected to: structure A→structure B→structure C to ensure that the contact order not only meets the timeline but also meets the spatial anatomical logic and is more medically reasonable.

[0125] S33, determining whether the operation sequence conforms to the surgical procedure template path based on an edit distance algorithm, to obtain an operation path discrimination report; the operation path discrimination report includes path integrity, path interruption, and path error.

[0126] The system compares the above generated operation sequence with the preset standard surgical procedure path, i.e., the template path. Illustratively, the edit distance (Edit Distance) algorithm is used as the matching degree evaluation index. Edit distance is a commonly used method to measure the difference between two sequences, defined as the minimum number of operations (operations include insertion, deletion, and replacement) required to convert one sequence into another.

[0127] Illustratively, suppose the template path is TEMPLATE=[structure A, structure B, structure C, structure D], and the operation sequence is ACTUAL=[structure A, structure C, structure E], the edit distance is to delete structure B (deletion operation), insert structure D (missing incomplete), and replace structure E with structure D (error operation). According to the size of the edit distance, the system can calculate the path deviation score, and output a structured operation path discrimination report accordingly. The report includes path integrity, path interruption, path error, and operation deviation score.

[0128] Specifically, if the template path requires the treatment of the uterine round ligament first, followed by the dissection of the uterine artery, and then the cleaning of the internal iliac lymph nodes, but the trainer directly jumps to the lymph node cleaning, the system will mark "path interruption: missing uterine artery treatment step". The generation of this report makes the surgical training not only focus on whether the task is completed, but also focus on whether the completed path is safe, standard, and reasonable, thereby realizing the landing of the path-dependent evaluation mechanism.

[0129] In one embodiment, the action trajectory sequence and the structure response state table are aligned according to the response time to obtain a list of false operation markers, including:

[0130] S41, performing surgical action filtering on the behavior type contained in the action trajectory sequence to obtain the tissue structure interaction action trajectory.

[0131] Illustratively, each unit in the action trajectory sequence records includes an operation number, an instrument motion trajectory, an identified behavior type (such as "contact", "push", "cut", "puncture", "peel"), and a start and end time interval of the behavior. The system performs a filtering according to the behavior type table, removes non-interaction actions (such as "approach", "hover"), and retains all interaction actions that have substantial contact with the anatomical model tissue structure (such as "cut", "push and pull"). The part of the action sequence after the filtering is the tissue structure interaction action trajectory.

[0132] S42, aligning the time sequence corresponding to the tissue structure interaction action trajectory with the response time to obtain an alignment result; the alignment result includes corresponding response state events and no corresponding response state events.

[0133] Each record in the structure response state table generally includes a tissue structure number, a response type (such as force change, electrical impedance mutation, breakpoint response), a response value, and a response occurrence time. The time interval of each interaction action is compared with the time point in the response state table, and a response time tolerance window Δt is set. If there is a response event within the time window, it is considered that the "action-response alignment is successful". Illustratively, the cutting action occurs at T3-T4; the response state of structure number A detects a break response at T3.1, and it is considered that the action is aligned with the structure response. The system labels all trajectories as having a response or no corresponding response in this way.

[0134] S43, if the alignment result is no corresponding response state event, mark the tissue structure interaction action trajectory as a false operation, and generate a false operation mark list.

[0135] Marking the non-response operation as a false operation generates a false operation mark list.

[0136] In one embodiment, based on the target tissue structure cleaning list, the surgical target hit result is generated according to the operation characteristics and the structure response state list, including:

[0137] S51, causally correlate the structure tracking matrix, the action trajectory sequence, and the response state according to a preset time window to obtain a cleaning object table.

[0138] The operation behavior is compared with the structural response event in a time window to determine whether there is a time and space overlap. For example, if a certain action produces obvious push-pull trajectories on the structure of region A within T1-T2, the structure A in region A exists at T0, the boundary is lost at T3, and the structure A produces deformation response at T1.3, it can be considered that the "action-structure-response" forms a causal chain, and the cleaning object item is obtained, and a cleaning object table is generated including the structure name, whether it is cut off, whether the response is consistent, and whether the structure tracking is supported.

[0139] S52, compare the cleaning object table with the target organizational structure cleaning list to obtain a cleaning object comparison table; the cleaning object comparison table includes whether it is a target organizational structure determination result.

[0140] The cleaning object table is compared with the target organizational structure cleaning list to obtain a cleaning object comparison table. Specifically, the actual cleaning object is compared with the preset target organizational structure cleaning list to determine whether the target structure is cleaned. The target organizational structure cleaning list is preset by the surgical task requirement in the form of structure number and region name. If the structure name of each cleaning object is mapped to the target list, it is recorded as "target hit", otherwise it is considered as "non-target cleaning" or "misoperation", and a cleaning object comparison table is generated.

[0141] S53, calculate the region corresponding according to the structure number and the response state according to the target organizational structure cleaning list to obtain the cleaning completion rate of each region.

[0142] According to the structure number and the response state, the region corresponding calculation is performed according to the target organizational structure cleaning list to obtain the cleaning completion rate of each region. For each target region structure number A i , the cleaning completion rate C i can be calculated as follows:

[0143]

[0144] Wherein, N 有效清扫响应 (A i ) is the number of structures in the region that successfully produce operation response; N 应清扫子结构 (A i ) is the total number of structures that should be cleaned in the region.

[0145] S54, generate a misoperation structure table according to the structure response state table and the target organizational result cleaning list.

[0146] Illustratively, all structures that are operated and produce responses but do not belong to the target organizational cleaning list are identified as "misoperation structures", and a misoperation structure table is generated.

[0147] It should be understood that although the steps in the flowcharts involved in the embodiments described above are shown in sequence according to the arrows, the steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, the execution of the steps is not strictly limited in sequence, and the steps can be executed in other sequences. Moreover, at least some of the steps in the flowcharts involved in the embodiments described above can include multiple steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of the steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least part of other steps or steps or stages in other steps.

[0148] Based on the same inventive concept, the embodiments of the present application also provide a gynecological pelvic-abdominal tumor surgery training result discrimination system for implementing the gynecological pelvic-abdominal tumor surgery training result discrimination method described above. The implementation scheme for solving the problem provided by the system is similar to the implementation scheme described in the above method, so the specific limitations in one or more gynecological pelvic-abdominal tumor surgery training result discrimination system embodiments provided below can refer to the limitations of the gynecological pelvic-abdominal tumor surgery training result discrimination described above, and will not be repeated here.

[0149] In one exemplary embodiment, as shown in Figure 3 a gynecological pelvic-abdominal tumor surgery training result discrimination system is provided, comprising:

[0150] an anatomical model, a server and a camera;

[0151] The anatomical model simulates the pelvic-abdominal structure of the human body, and is used to provide a pelvic-abdominal tumor surgery training for a training personnel, and the anatomical model is integrated with a response module for collecting a surgery process response signal stream;

[0152] The response module includes a sensor and a communication node; the response module is connected with the server through a built-in first data transmission module;

[0153] The camera is connected with the server through a built-in second data transmission module;

[0154] The server is used to implement the steps of the gynecological pelvic-abdominal tumor surgery training result discrimination method.

[0155] The anatomical model is used to simulate the real human pelvic and abdominal cavity anatomical structure for the trainer to perform actual operation. The anatomical model is internally provided with a plurality of structure partitions corresponding to key anatomical sites such as uterus, oviduct, ovary, pelvic lymphatic region, abdominal aortic side, etc. Each tissue region is distinguished by different color or material for image recognition and operation training. The tissue hierarchy is supported for partition shaping, so that the cutting, separation and other operations have a realistic feeling. The anatomical model provides a physical environment for surgical training operation, which carries a response module, records the force or state change of each structure in the training, is the generation source of the structure tracking matrix and response state.

[0156] The response module integrated in the anatomical model is used to collect the tissue reaction signal in the operation process in real time, to ensure the real feedback to the training behavior, including pressure sensor, tension sensor, electrical impedance sensor, etc., which are arranged on the key structure points of the model. The communication node is a signal communication node of the detachable anatomical tissue, such as the disconnection of the uterus and the anatomical model in the hysterectomy operation.

[0157] The response module is connected with the server through the built-in first data transmission module, which can adopt wired or wireless communication mode.

[0158] The camera adopts fixed or track type high-definition camera, which covers the whole operation area. The camera is connected with the server through the built-in second data transmission module; USB3.0, wired network port or wireless video transmission module can be adopted;

[0159] The server has high concurrent data writing ability and high storage capacity, supports long time training video and response data storage, to realize each step of the gynecological pelvic and abdominal cavity tumor surgery training result discrimination method.

[0160] Based on the same inventive concept, the embodiment of the present application also provides an anatomical model for gynecological pelvic and abdominal cavity tumor surgery training, which is used to realize the gynecological pelvic and abdominal cavity tumor surgery training result discrimination method.

[0161] In an exemplary embodiment, as shown in Figures 4-6 A gynecological pelvic and abdominal cavity tumor surgery training anatomical model is provided, comprising:

[0162] The main part simulating the pelvic and abdominal cavity structure is used for training, which widely includes the instrument assembly of hysterectomy, pelvic lymph node dissection and abdominal aortic side lymph node dissection operation, the response module integrated in the main part and the data transmission unit. The response module includes sensor assembly and communication node.

[0163] The main body part has a simulation tissue module simulating human skin, pelvic tissue structure, abdominal aorta and inferior vena cava, ureter, uterus and its accessories, multiple organs in the upper abdomen and lymph node groups; the lymph node groups are attached to the pelvic cavity and abdominal aorta and inferior vena cava; each simulation tissue module of the main body part has different material rigidity to simulate the real human tissue morphology, texture, separability and adhesion degree;

[0164] The sensor assembly comprises a pressure sensor for detecting the operating force acting on the simulation tissue module in the cutting, pulling and cleaning operation, and the pressure sensor is arranged at the attachment point of the lymph node group, the blood vessel wall and the anatomical interface position;

[0165] The communication node comprises a communication node arranged on each anatomical interface to obtain the signal on the anatomical interface and obtain the response signal of each simulation tissue module in the surgical training process.

[0166] The data transmission unit is connected to the response module and is used for transmitting the data collected by the response module to the server, and the server is used for implementing the steps of the gynecological pelvic-abdominal cavity tumor surgery training result judgment method.

[0167] The anatomical model is designed to cooperate with the training system to complete the training process of complex gynecological surgical operations including hysterectomy, pelvic lymph node cleaning and para-aortic lymph node cleaning, and to support the process judgment and scoring analysis of the surgical training results. The main body part is constructed by a plurality of simulation tissue modules, and each simulation tissue module is constructed by a high-molecular biomimetic material with different elastic modulus, so as to reproduce the real morphology, touch, operability, separability and physiological adhesion characteristics of each tissue. Especially at the attachment site of the lymph node group, the model accurately simulates the common adhesion structure in the pelvic cavity and the para-aortic artery by partition bonding technology, which is used to train the accuracy of the cleaning area and the separation feeling.

[0168] The sensor assembly is arranged at the key structure position of the main body part, and comprises a plurality of pressure sensors arranged on the blood vessel wall, the lymph node attachment point and the anatomical interface of each tissue. The sensor is used to detect the acting force of different types of operations such as cutting, pulling and cleaning during the training operation in real time, and form a time-stamped operation force data stream, which provides a basis for subsequent response state recognition and operation intensity analysis.

[0169] The communication node is mainly installed on a plurality of anatomical interfaces, and is used to collect the electrical contact state or magnetic disconnection state during the separation of tissues, so as to judge whether the adjacent tissue structure is cut off, cleaned or removed. The communication node converts the physical state change at the anatomical interface into a standardized response signal through an embedded logic circuit, and uploads the signal to the server in real time.

[0170] Through the above structural configuration, the anatomical model not only supports high-fidelity simulation of surgical training operations, but also provides accurate data basis for subsequent structure tracking, action behavior identification, and operation rationality determination.

[0171] In an embodiment, a computer device is provided, comprising a memory and a processor, the memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0172] In an embodiment, a computer readable storage medium is provided, which stores a computer program, and the computer program is executed by a processor to implement the steps in the above method embodiments.

[0173] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts are described in the method embodiments. The above described device embodiments are only illustrative, and the components described as separate components can or can not be physically separated, and the components displayed as units can or can not be physical units, i.e., they can be located in one place or distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the present disclosure according to actual needs. Those skilled in the art can understand and implement it without creative labor.

[0174] The above described embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the patent scope of the application. It should be pointed out that, for those skilled in the art, without departing from the concept of the present application, several modifications and improvements can be made, which are all within the protection scope of the present application.

Claims

1. A method for judging the result of gynecological pelvic and abdominal cavity tumor surgery training, applied to a server in a gynecological pelvic and abdominal cavity tumor surgery training result judgment system, characterized in that, The gynecological pelvic and abdominal tumor surgery training result discrimination system is connected with an anatomical model and a camera; the anatomical model simulates the pelvic and abdominal cavity structure of a human body and is used to provide a training personnel for pelvic and abdominal tumor surgery training, and the anatomical model is integrated with a response module for collecting a surgery process response signal stream; the response module comprises a sensor assembly and a communication node; the response module is connected with the server through a built-in first data transmission module; The camera is connected with the server through a built-in second transmission module; The gynecological pelvic and abdominal tumor surgery training result discrimination method comprises the following steps: Obtaining video data and a response signal list in the surgery training process; the response signal list comprises response signal streams of a plurality of structure nodes of the anatomical model; Extracting data features of the video data to obtain operation features; the operation features comprise a structure tracking matrix and a motion trajectory sequence; Performing time sequence processing on the response signal streams of the response signal list to obtain a structure response state table; the structure response state table comprises structure numbers, response states and response times; Generating a surgery process judgment report according to the operation features and the structure response state table; Based on a preset scoring table, scoring according to the surgery process judgment report to obtain a surgery training result score.

2. The method of claim 1, wherein, The operation features are extracted from the video data, comprising the following steps: Performing key frame extraction on the video data to obtain an anatomical model image sequence and a motion occurrence frame image sequence; Using a color segmentation algorithm to identify boundaries corresponding to initial states and final states of tissue structures of the anatomical model image sequence to obtain a structure tracking matrix; the structure tracking matrix comprises tissue structure boundaries, corresponding change features and time stamps; Performing surgery instrument detection on the motion occurrence frame image sequence based on optical flow analysis to obtain an instrument motion trajectory; Aligning the instrument motion trajectory with the structure tracking matrix according to the time stamps and performing behavior type identification to obtain a motion trajectory sequence; the motion trajectory sequence comprises kinematic trajectories and behavior types; The result tracking matrix and the motion trajectory sequence are determined as the operation features.

3. The method of claim 2, wherein, The surgery process judgment report is generated according to the operation features and the structure response state table, comprising the following steps: Reconstructing a tissue structure contact sequence according to the structure tracking matrix to obtain an operation path discrimination report; Aligning the motion trajectory sequence and the structure response state table according to the response times to obtain a false operation marker list; Based on a target tissue structure cleaning list, generating a surgery target hit result according to the operation features and the structure response state list; the surgery target hit result comprises a cleaning object comparison table, a cleaning completion rate and a misoperation structure table; Generating the surgery process judgment report according to the operation path discrimination report, the false operation marker list and the surgery target hit result.

4. The method of claim 3, wherein, The operation path discrimination report is obtained by reconstructing a tissue structure contact sequence according to the structure tracking matrix, comprising the following steps: Arranging the tissue structure boundaries corresponding to the change features in a sequence of sequentially advancing according to the time stamps to obtain a tissue structure change sequence; reconstructing the tissue structure contact sequence according to the tissue structure change sequence, to obtain a surgical procedure operation sequence; determining whether the surgical procedure operation sequence conforms to a surgical procedure template path based on an edit distance algorithm, to obtain an operation path discrimination report; the operation path discrimination report includes path completeness, path interruption, and path error.

5. The method of claim 3, wherein, aligning the action trajectory sequence and the structure response state table according to the response time, to obtain a false operation marking list, including: performing surgical action screening on the behavior type contained in the action trajectory sequence, to obtain a tissue structure interaction action trajectory; aligning a time sequence corresponding to the tissue structure interaction action trajectory with the response time, to obtain an alignment result; the alignment result includes a corresponding response state event and a non-corresponding response state event; if the alignment result is the non-corresponding response state event, marking the tissue structure interaction action trajectory as a false operation, and generating a false operation marking list.

6. The method of claim 3, wherein, generating a surgical target hit result according to the operation feature and the structure response state table based on the target tissue structure cleaning list, including: performing causal correlation on the structure tracking matrix, the action trajectory sequence, and the response state according to a preset time window, to obtain a cleaning object table; comparing the cleaning object table with the target tissue structure cleaning list, to obtain a cleaning object comparison table; the cleaning object comparison table includes a target tissue structure determination result; performing region corresponding calculation according to the structure number and the response state according to the target tissue structure cleaning list, to obtain a cleaning completion rate corresponding to each region; generating a misoperation structure table according to the structure response state table and the target tissue result cleaning list.

7. A system for discriminating results of gynecological pelvic-abdominal tumor surgery training, characterized by comprising: a gynecological pelvic-abdominal tumor surgery training result discriminating device; and a gynecological pelvic-abdominal tumor surgery training result discriminating program. The system comprises an anatomical model, a server, and a camera. The anatomical model simulates the pelvic and abdominal cavity structure of the human body, and is used to provide training personnel with pelvic and abdominal tumor surgery training. The anatomical model is integrated with a response module for collecting surgical procedure response signal streams. The response module comprises sensors and communication nodes. The response module is connected with the server through a built-in first data transmission module. The camera is connected with the server through a built-in second data transmission module. The server is used to implement the steps of the method of any one of claims 1-6.

8. An anatomical model for gynecologic pelvic-abdominal tumor surgery training, characterized by, The system comprises an anatomical model, a server, and a camera. The main body part simulates the pelvic and abdominal cavity structure, and is used to train an instrument assembly widely containing hysterectomy, pelvic lymph node cleaning, and abdominal aortic lymph node cleaning operation. The response module and the data transmission unit are integrated inside the main body part. The response module comprises a sensor assembly and a communication node. The main body part has simulation tissue modules simulating human skin, pelvic tissue structure, abdominal aorta and inferior vena cava, ureter, uterus and its appendages, upper abdominal organs, and lymph node groups. The lymph node groups are attached to the pelvic cavity, the abdominal aorta, and the inferior vena cava. Each simulation tissue module of the main body part has different material rigidity to simulate the real human tissue morphology, texture, separability, and adhesion degree. The sensor assembly comprises a pressure sensor for detecting the operating force acting on the simulated tissue module in the resection, traction and cleaning operation, and the pressure sensor is arranged at the lymph node group attachment point, the blood vessel wall and the anatomical interface position. The communication node comprises a device for detecting whether the simulated tissue module including the uterus and its appendages, the upper abdominal multi-organ and the abdominal tissue structure is resected and separated at the anatomical interface; the communication node is arranged at each anatomical interface to obtain the signal on the anatomical interface and obtain the response signal of each simulated tissue module in the surgical training process. The data transmission unit is connected to the response module and is used for transmitting the data collected by the response module to a server, and the server is used for implementing the steps of the method in any one of claims 1-6. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The processor executes the computer program to implement the steps of the method in any one of claims 1-6.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method in any one of claims 1-6.

Images