A method and system for recognizing contributions to laparoscopic state transitions and evaluating skills.

By identifying the movement of instruments and the state of targets during laparoscopic training, and combining this with task state transition rules, the preservation and utilization of preconditions for operation are verified. This solves the problem of misjudgment in kinematic evaluation in existing technologies and achieves more accurate recognition of movement behavior and skill evaluation.

CN123070795APending Publication Date: 2026-11-10MEDI RUIER INTERNATIONAL INTELLIGENT TECHNOLOGY (BEIJING) CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611252413.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-18
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

In current evaluations of laparoscopic physical training, it is difficult to identify the actual task contribution of preparatory movements based solely on the amount of exercise or changes in the target object. Furthermore, kinematic evaluations are prone to misjudging necessary preparatory movements as inefficient or movements that are not utilized by subsequent effective operations as high-quality movements.

Method used

By acquiring instrument motion data and physical training target state data during laparoscopic training, motion segments are divided according to preset task state transition rules. It is also verified whether the preconditions for operation formed by the preceding motion segments are maintained and/or actually utilized in subsequent movements. Combined with whether the target state transition is completed in subsequent movements, the contribution type of the motion segment is determined.

Benefits of technology

It improves the accuracy and interpretability of laparoscopic physical training motion behavior recognition, reduces the probability of misjudging necessary preparatory movements as redundant movements, and enhances the reliability and repeatability of evaluation results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123070795A_ABST
    Figure CN123070795A_ABST
Patent Text Reader

Abstract

This invention discloses a method and system for recognizing contributions to laparoscopic state transitions and evaluating skills. It acquires motion data of the training equipment and state data of the training target during physical training, and divides continuous motion into multiple motion segments according to task state transition rules. For the first motion segment that does not directly trigger a target state transition, the preconditions for its formation are determined, and it is verified whether the subsequent second motion segment maintains and / or utilizes these preconditions and actually completes the target state transition; if satisfied, the first motion segment is determined to have an enabling contribution to the target state transition. Skill evaluation results are generated based on the contribution relationship between motion segments and task state transitions. This invention can reduce misclassifications that occur when automatically classifying physical training motion behaviors solely based on kinematic indicators, improving the reliability and interpretability of motion behavior recognition.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the fields of medical skills training, computer-aided training evaluation, motion data processing, and sensor information processing. In particular, it relates to a method and system for establishing a contribution relationship between motion segments of training equipment and specific task state transitions during laparoscopic physical training, and for identifying motion behavior and evaluating skills accordingly. Background Technology

[0002] Laparoscopic skills training typically requires trainees to use laparoscopic instruments within a physical training chamber to complete standardized training tasks such as object transfer, graphic cutting, and suturing. To improve the objectivity of training evaluation, existing technologies can collect data such as instrument position, posture, speed, acceleration, and motion path, and calculate indicators such as completion time, path length, motion smoothness, and workspace based on this data.

[0003] However, evaluating a movement solely based on kinematic metrics such as path, speed, or smoothness cannot fully reflect its actual role in a specific task. For example, to achieve effective grasping, obstacle avoidance, needle posture adjustment, or restoration of the correct operating space, the operator may need to perform certain retractions, rotations, or positional adjustments. While such movements do not immediately change the task state of the training target, they may establish the necessary spatial, posture, or interaction conditions for subsequent effective operations.

[0004] Conversely, some movements, even if they have short paths or high speeds, should not be considered high-quality movements solely based on their kinematic performance if they are not utilized by subsequent effective operations, do not drive task state transitions, and do not complete error recovery.

[0005] Therefore, a technical solution is needed that can determine, based on the state transition process of an entity training task, whether a motion segment that does not directly trigger a task state transition is a prerequisite for the subsequent actual completion of the state transition and provides an operational condition that can be actually utilized, thereby distinguishing between necessary preparatory motions that contribute to the task and redundant motions that do not contribute to the task. Summary of the Invention

[0006] I. Technical problems to be solved

[0007] This invention aims to solve the problem in existing laparoscopic physical training evaluations that it is difficult to identify the actual task contribution of preparatory movements based solely on the amount of exercise or real-time changes in the target object.

[0008] Furthermore, the present invention aims to address the problem that similar machine motion segments may have different operational meanings in different task contexts, and that existing kinematic evaluations are prone to misjudging necessary preparatory movements as inefficient movements, or misjudging movements that are not utilized by subsequent effective operations as effective movements.

[0009] Furthermore, the present invention aims to correlate physical measurement data from physical training equipment and physical training targets, so as to establish a detectable and repeatable correspondence between the automatic identification results of motion segments and the state transition of physical tasks, thereby reducing misclassification when automatically classifying physical training motion behavior based solely on kinematic indicators.

[0010] II. Technical Solution

[0011] To address the aforementioned technical problems, this invention provides a method for recognizing and evaluating movement behaviors during laparoscopic training. This method acquires motion data of the training apparatus and state data of the physical training target during laparoscopic training; determines a first task state, a second task state, and a target state transition from the first task state to the second task state based on preset task state transition rules; and divides continuous motion into multiple motion segments based on motion data, apparatus-target interaction events, and / or task state change events.

[0012] For a first motion segment that does not directly trigger the target state transition, at least one detectable precondition for the formation of the first motion segment is determined based on the motion state of the device corresponding to the motion segment, the relative state between the device and the physical training target, and / or the state of the physical training target.

[0013] Further identify a second motion segment that occurs after the first motion segment and is associated with the target state transition, detect whether the operation precondition remains valid and / or is actually utilized by the second motion segment during the execution of the second motion segment, and detect whether the second motion segment actually enables the training task to complete the target state transition.

[0014] If the preconditions for operation are maintained and / or utilized by the second motion segment and the second motion segment actually completes the target state transition, the first motion segment is identified as the motion segment that enables the target state transition; and a laparoscopic skill evaluation result is generated based on the state transition contribution relationship between at least one motion segment and at least one task state transition.

[0015] The preconditions for operation are state conditions formed by the first motion segment that can be detected by the system and maintained or utilized in subsequent operations, including one or more of the following: instrument position conditions, instrument posture conditions, instrument axial rotation conditions, instrument and target relative position conditions, workspace or safety clearance conditions, gripping or contact conditions, target position or posture conditions, and target tension or exposure conditions.

[0016] The phrase "utilized by subsequent motion segments" means that during the execution of the second motion segment, the second motion segment satisfies or maintains the preconditions for the operation formed by the first motion segment, and the second motion segment interacts with the same training device, the same physical training target, or the training object associated with the preconditions for the operation, thereby actually completing the target state transition corresponding to the preconditions for the operation.

[0017] In one implementation, the system can also identify direct contributing motion segments that directly cause the target state transition, recovery contributing motion segments that restore the erroneous state to a legal and continueable state, and non-contributing motion segments that do not directly complete the target state transition, do not establish operational preconditions that are actually used by subsequent target state transitions, and do not complete error recovery.

[0018] In one implementation, a target state transition can be associated with multiple motion segments; wherein at least one motion segment can serve as a directly contributing motion segment, and at least one preceding motion segment can serve as an enabling contributing motion segment. The association between motion segments and state transitions can be determined based on one or more of the following: time window, operation sequence window, instrument identification, target identification, spatial relationship continuity, posture continuity, and interaction continuity.

[0019] The motion data of the training equipment can be obtained through optical vision, passive or active visual markers, RGB-D depth vision, electromagnetic positioning, mechanical coding, inertial sensing, robot encoders, or multi-sensor fusion. This invention is not necessarily limited to specific motion data acquisition hardware.

[0020] The state data of the target object in the training can be obtained from RGB images, depth images, visual markers, object detection, image segmentation, contact sensing, force sensing, position sensing, switch quantity sensing, or multi-source fusion.

[0021] Task state transition rules can be represented by state diagrams, state tables, finite state machines, rule sets, probabilistic state models, or equivalent data structures. Preferably, the determination of task states and the assessment of state transition contributions can be implemented by deterministic rules or finite state machines, thus allowing the invention to be implemented without relying on a specific artificial intelligence model.

[0022] When a machine learning model is used to assist in task state recognition or interactive event recognition, its input includes at least instrument motion data, target state data, and / or instrument-target relative state data, and its output includes at least one of task state, interactive event, or precondition state; the machine learning model is not a necessary limitation of the independent claims of this invention.

[0023] In a preferred embodiment, the laparoscopic physical training system includes a physical training box, at least one laparoscopic training instrument, at least one physical training target, at least one sensor subsystem, and a data processing device communicatively connected to the sensor subsystem. The sensor subsystem acquires sensing signals characterizing the movement of the training instrument and / or the physical state of the physical training target, and establishes time information for the sensing signals. The data processing device obtains the state data of the training instrument and the physical training target based on the sensing signals, and performs task state transition contribution identification based on the state data.

[0024] In one implementation, motion data of training equipment and state data of physical training targets from different sensor sources are converted to a common task timeline; when spatial correlation is required, the data can also be converted to a common coordinate reference system. This allows for the establishment of machine-processable correlation records between motion segments and target state transitions based on the start and end times of the motion segment, equipment identification, target identification, and corresponding task state.

[0025] In one implementation, the contribution attribution module generates a machine-readable contribution record for each evaluated motion segment. The contribution record includes at least one or more of the following: motion segment identifier, associated task state transition identifier, training equipment identifier, physical training target identifier, precondition state, and contribution type. The contribution record can be stored, transmitted, used to generate training device output, or as input for subsequent training evaluation processing.

[0026] III. Beneficial Effects

[0027] Compared to schemes that evaluate solely based on path length, speed, acceleration, or motion smoothness, this invention attributes the contribution of preparatory motion to state transition by verifying whether the operational preconditions formed by the preceding motion segment are maintained and / or actually utilized in subsequent motion, and by combining this with whether the target state transition is actually completed.

[0028] Therefore, the present invention can reduce the probability of misjudging necessary preparation, withdrawal, adjustment or posture adjustment movements as redundant movements, and at the same time reduce the probability of movements that do not contribute to the actual task being misjudged as high-quality movements due to short paths or high speeds.

[0029] This invention can also locate the evaluation results to the specific task state transition stage and output motion behavior information with task context, such as direct contribution, enabling contribution, recovery contribution and non-contribution, thereby improving the accuracy, repeatability and interpretability of motion behavior recognition in laparoscopic physical training.

[0030] Furthermore, by performing time-related and task-state-related correlations on physical measurement data from physical training equipment and physical training targets, and verifying whether the physical preconditions formed by the preceding motion are actually maintained and / or utilized by subsequent physical operations, this invention can reduce misclassifications that occur when automatically classifying physical motion segments based solely on path length, speed, or instantaneous state changes, thereby improving the reliability of automatic detection and classification results of physical training motion behaviors. Attached Figure Description

[0031] Figure 1 This is a schematic diagram of the overall structure of the laparoscopic motion behavior recognition and skill evaluation system of the present invention;

[0032] Figure 2 This is a schematic diagram of the overall process of the method of the present invention;

[0033] Figure 3 This is a schematic diagram of the task states and state transition rules of the present invention;

[0034] Figure 4 A schematic diagram of the state transition contribution attribution process for this invention;

[0035] Figure 5 A schematic diagram illustrating the different contribution attributions of similar motion segments under conditions where the preconditions for operation are actually utilized or not.

[0036] Figure 6 A schematic diagram illustrating an example of state transitions and contribution attribution in an object transfer task;

[0037] Figure 7 A schematic diagram illustrating the state transitions and contribution attribution in a graphics cutting task;

[0038] Figure 8 A schematic diagram illustrating state transitions and contribution attribution in a stitching task;

[0039] Figure 9 This is a schematic diagram of the physical sensing data and state transition contribution processing chain of the present invention. Detailed Implementation

[0040] The present invention will be further described below with reference to the accompanying drawings and embodiments. It should be understood that the following embodiments are used to illustrate the technical solutions of the present invention, and are not intended to limit the scope of protection of the present invention. Without departing from the core concept of the present invention, those skilled in the art can make equivalent substitutions for the data acquisition method, state representation method, motion segmentation method, and specific calculation method of contribution relationship.

[0041] like Figure 1As shown, the system may include a motion data acquisition module 101, a target state acquisition module 102, a task state transition module 103, a task rule base 104, a motion segmentation module 105, an interaction event detection module 106, a precondition determination module 107, a state transition detection module 108, a contribution attribution module 109, and a skill evaluation module 110. Each module can be implemented by a software program on the same processor, or it can be implemented collaboratively by multiple local or remote computing devices.

[0042] Implementation Method 1: General State Transition Contribution Attribution Method

[0043] like Figure 2 As shown, in step S201, motion data of the training device and state data of the physical training target are acquired. The motion data of the training device can be represented as I(t) = {p(t), q(t), g(t)}, where p(t) represents the spatial position of the device's end effector, q(t) represents the device's posture or axial rotation state, and g(t) represents the opening / closing or gripping state of the device's end effector. All of the above data items can be acquired, or a portion can be acquired depending on the specific training task.

[0044] The state data of the entity training target can be represented as O(t), which at least represents one of the following: position, orientation, boundary, grasping state, transfer state, cutting state, stitching state, integrity, or whether it is in the target region.

[0045] In step S202, a task state set S={S0,S1,…,Sn,SE} and a state transition set E are pre-established according to the task rule base 104. Figure 3 For example, S0 is the initial state of the target object, S1 is the state that meets the preconditions for operation, S2 is the state where a valid interaction occurs, S3 is the target state completion state, and SE is the error state. State transitions E01, E12, and E23 describe the normal task flow. Error events can cause the task to enter SE, and recovery state transitions return the task from SE to a valid and continueable state.

[0046] In step S203, based on at least one of the following: change in device motion state, change in device motion direction, change in device posture, device entering or leaving the target interaction area, device-target interaction event, change in target state, and task state transition event, continuous motion is divided into motion segments M1, M2, ..., Mm.

[0047] In steps S204 to S205, the first motion segment Mi that does not directly trigger the state transition Eab of the target to be evaluated is determined from multiple motion segments, and the precondition P for the formation of Mi is determined. P is a detectable state condition, such as the instrument end entering a predetermined spatial region, the instrument rotating axially into a predetermined posture range, the instrument and the target forming a predetermined relative position, the physical target maintaining a specific grasping state, and the training material forming a tension or exposure state that satisfies subsequent cutting, etc.

[0048] In steps S206 and S207, a second motion segment Mj that occurs after Mi and is associated with Eab is selected. The system verifies whether P remains valid at the beginning or during the execution of Mj, and verifies whether Mj completes the corresponding instrument-target interaction using the spatial, posture, grip, contact, tension, exposure, or other state conditions represented by P. The system also detects whether Eab is actually completed.

[0049] In one feasible determination method, if Mi and Mj target the same instrument or a task-related instrument, act on the same target or a task-related target, P is still true at the start of Mj or at a critical interaction moment of Mj, and Eab is completed within a predetermined operation sequence window, then it is determined that Mj maintains and / or utilizes P. The predetermined operation sequence window can be a fixed time window or a range of subsequent motion segments determined by the task state machine.

[0050] In a preferred determination method, for the candidate operation precondition Pk formed by the first motion segment Mi, the system records the state value of Pk at the end of Mi, and detects at least one second motion segment Mj within the subsequent operation sequence window associated with the target state transition Eab. If Pk still meets the preset conditions at the beginning of Mj and / or when Mj undergoes a critical instrument-target interaction, Mj acts on the same or task-related training instrument and physical training target associated with Pk, and Eab is detected as actually completed within the operation sequence window, then it is determined that Pk is maintained and / or actually utilized by Mj, and Mi is identified as having an enabling contribution to Eab.

[0051] Conversely, when the candidate operation precondition has expired before subsequent key interactions, the subsequent motion acts on an object that is not related to the candidate operation precondition, or the target state transition is completed through another set of state conditions that are unrelated to the candidate operation precondition, the first motion segment is not attributed to the enabling contribution of the target state transition based on the candidate operation precondition.

[0052] In step S208, if P is maintained and / or utilized by Mj and Mj actually completes Eab, then Mi is determined to have an enabling contribution to Eab. If a motion segment directly triggers Eab, it can be determined as a direct contribution; if the task first enters the error state SE, and subsequent motion restores the task to the legal state, then the corresponding motion segment can be determined as a recovery contribution; if a motion segment does not form a precondition that is actually utilized by the subsequent target state transition, does not directly complete the state transition, and does not complete error recovery, then it can be determined as a non-contributing motion segment.

[0053] The skill evaluation module 110 generates evaluation results based on the contribution relationship of state transitions. The evaluation results may include the direct contribution amount of motion, enabling contribution amount of motion, recovery contribution amount of motion, and non-contribution amount of motion corresponding to each task state transition, as well as the time, path length, number of occurrences, proportion, or level of the above-mentioned motion amounts. The evaluation results may also output textual explanations, such as "The current additional motion mainly occurs before the target state transition E12, and some of the adjustments were not effectively utilized by subsequent interactions."

[0054] Implementation Method 2: Object Transfer Training Task

[0055] like Figure 6 As shown, in an object transfer task, the task state can be defined as follows: S0, the target object is at the starting position; S1, the target object has been effectively grasped; S2, the target object has completed the predetermined spatial transfer; S3, the target object has been correctly released at the target position.

[0056] Motion segment M1 is used for the instrument to approach the target object, M2 is used to adjust the gripping angle, and M3 completes the gripping and directly triggers the state transition from S0 to S1. M4 can be the retraction obstacle avoidance motion after gripping. It does not directly complete the state transition from S1 to S2, but it forms a safety gap P. If M5 subsequently completes the target object transfer and actually triggers the transition from S1 to S2 while maintaining this safety gap, then M4 is determined to have an enabling contribution to the transition from S1 to S2.

[0057] As a clearly defined preferred retreat combination, when the safety gap P formed by the first motion segment is maintained in the critical stage of the subsequent transfer motion, and the second motion segment acts on the same or task-related entity training target, and the second motion segment actually completes the target spatial transfer state transition from S1 to S2, the first motion segment is determined to have an enabling contribution to S1 to S2.

[0058] If there exists a motion M4a that is similar to the spatial trajectory of M4, but the candidate safety gap formed by M4a is not maintained or actually utilized in subsequent motions, and the corresponding state transition is not completed subsequently, then M4a does not automatically obtain an enabling contribution because of its similarity to the trajectory of M4, and can be further classified as a non-contribution adjustment or an invalid round trip.

[0059] Implementation Method 3: Graphic Cutting Training Task

[0060] like Figure 7 As shown, the physical training target for the graphic cutting task can be training material with a preset target contour. The target object's state data can include the completed cutting contour, the current cutting point, the target contour deviation, the material orientation, and the material tension or exposure state.

[0061] The first motion segment M1 can be achieved by using an auxiliary device to traction or adjust the posture of the training material, creating tension and / or cutting angle conditions P that satisfy subsequent effective cutting. If the subsequent second motion segment M2 actually completes a new effective cutting segment while P remains valid, then M1 is determined to have an enabling contribution to the corresponding cutting state transition. Therefore, the system will not classify M1 as an invalid motion simply because it increases the device path.

[0062] As a set of clearly defined preferred retreat combinations, the first motion segment is identified as the enabling contributing motion segment for the corresponding cutting state transition when the material tension and / or exposure state P formed by the first motion segment still holds when the second motion segment undergoes a critical interaction of effective cutting, and when the second motion segment actually forms a new effective cutting segment.

[0063] If the candidate conditions formed by the traction or posture adjustment motion are not actually used by the subsequent cutting motion, or if the training material is subsequently restored to another state and cut by different conditions, then the preceding motion does not gain an enabling contribution to the transition of the cutting state.

[0064] Implementation Method 4: Suturing Training Task

[0065] like Figure 8 As shown, the suturing task can be defined as follows: S0 is the state where the needle is not in place, S1 is the state where the needle posture meets the needle insertion condition, S2 is the state where the needle insertion is completed, and S3 is the state where the needle is withdrawn.

[0066] The first motion segment M1 can be needle rotation or instrument posture adjustment. M1 itself does not directly complete needle insertion, but it brings the needle tip position, needle body direction, and / or the posture of the needle relative to the training material into the needle insertion condition P. If the subsequent second motion segment M2 completes needle insertion and triggers the state transition from S1 to S2 while the posture condition is maintained or actually utilized, then M1 is determined to have an enabling contribution to S1 to S2.

[0067] As a clearly defined preferred defensive combination, when the needle posture condition P formed by the first motion segment is still valid at the critical moment of needle insertion in the second motion segment, and the second motion segment actually completes the needle insertion state transition from S1 to S2 using the needle posture, the first motion segment is determined to have an enabling contribution to S1 to S2.

[0068] If the operator abandons the posture condition after M1, changes the needle direction again, and completes needle insertion through a new posture unrelated to P, then M1 is not determined as an enabling contribution to the needle insertion state transition. This determination is used to avoid evaluating motion contributions solely based on trajectory similarity or preset expert action templates.

[0069] Implementation Method 5: Error Recovery and Non-Contribution Determination

[0070] In one implementation, the task state machine sets an error state SE. When the system detects that the task has entered the error state SE from a legal task state based on the target state and / or the device-target interaction event, it establishes an error state transition record. If the subsequent recovery motion segment enables the physical training target and / or device to meet the preset legal and continueable state conditions again, and a recovery state transition from SE back to a legal state is detected, the corresponding recovery motion segment is determined to have a recovery contribution.

[0071] For movement segments that neither directly complete the target state transition, nor form the operational prerequisites that can be actually used by subsequent movements, and which have not completed the process of returning from an error state to a legal and continueable state, they can be identified as non-contributing movement segments or further subdivided into behavior types such as invalid round trips, searches, and repeated adjustments, according to the task rules.

[0072] Implementation Method Six: Physical Sensor Data Processing Chain and Deployment Method

[0073] like Figure 9 As shown, in a preferred embodiment, the physical states of the physical training device 901 and the physical training target 902 are collected by the sensor subsystem 903. The data processing device 904 receives the sensor signals, and the time synchronization and state extraction unit 905 generates training device state data and target state data with a common task time axis. Combined with the task state transition rule 104, the precondition detection and actual utilization verification unit 906 performs state transition contribution identification. The contribution attribution result can form a state transition contribution record 907, which is provided to the skill evaluation module or training equipment output 908.

[0074] In a preferred optical embodiment, the motion state of the training device is obtained through one or more fixed optical sensors, and the state of the physical training target is obtained through the same optical sensor or another target state sensor. The data processing device obtains the device position and / or attitude and the target state according to calibration parameters, and correlates the device state with the target state through common time information. This embodiment is only one possible implementation route and does not limit the invention to the use of specific cameras, visual markers, or visual algorithms.

[0075] In another embodiment, the motion data of the training equipment can be obtained by mechanical coding, electromagnetic positioning, inertial sensing, robot encoders, or a combination thereof, and the state of the physical training target can be obtained by position, contact, force, switching quantity, or other physical state sensors. As long as the corresponding data can form the equipment state and target state associated with the task state transition rules, the same state transition contribution recognition process can be performed.

[0076] The motion data acquisition, entity target state detection, and state transition contribution attribution of the present invention can be performed on the local processor of the training device, a computer connected to the training device, an edge computing device, or a remote server; or a combined architecture of local acquisition and remote evaluation can be adopted.

[0077] This invention can perform retrospective analysis of several preceding motion segments and determine enabling contributions after the target state transition is completed, or it can perform a unified analysis of the complete motion sequence after training is completed. Real-time evaluation is not a necessary limitation of this invention.

[0078] Without altering the core technical concept of this invention, the contribution of motion segments to state transitions can be represented by discrete contribution types or continuous contribution degrees. Continuous contribution degrees can be calculated based on one or more of the following: the degree of preservation of preconditions, the degree of object association, the continuity of interaction, and the degree of completion of state transitions; however, specific scoring weights do not constitute a necessary limitation for the implementation of this invention.

[0079] The laparoscopic physical training task described in this invention is used for skill training and evaluation of training equipment in a simulated training environment. The physical training target is a training model or training material. This invention does not require the performance of surgical treatments or diagnostic procedures on a human body.

Claims

1. A method for recognizing motor behavior and evaluating skills in a laparoscopic physical training task, characterized in that, include: Acquire motion data of the training equipment and state data of the physical training target during laparoscopic physical training; The first task state, the second task state, and the target state transition from the first task state to the second task state are determined according to the preset task state transition rules. The continuous motion of the training device is divided into multiple motion segments according to the motion data, device-target interaction events, and / or task state change events. At least one first motion segment that does not directly trigger the target state transition is determined from the multiple motion segments. At least one operational precondition for the formation of the first motion segment is determined according to the device motion state corresponding to the first motion segment, the relative state between the device and the physical training target, and / or the state of the physical training target. Identify at least one second motion segment that occurs after the first motion segment and is associated with the target state transition; detect whether the operational preconditions are maintained and / or utilized during the execution of the second motion segment; and detect whether the second motion segment actually enables the training task to complete the target state transition. If the operational preconditions are maintained and / or utilized by the second motion segment, and the second motion segment actually completes the target state transition, identify the first motion segment as a motion segment that enables the target state transition. Generate a laparoscopic skill evaluation result based on the state transition contribution relationship between at least one motion segment and at least one task state transition.

2. The method according to claim 1, characterized in that, The preconditions for operation include at least one of the following: spatial position conditions of the training device, posture conditions of the training device, axial rotation conditions of the training device, relative position conditions between the training device and the physical training target, workspace or safety clearance conditions, gripping or contact conditions, position conditions of the physical training target, posture conditions of the physical training target, tension conditions of the physical training target, and exposure conditions of the physical training target.

3. The method according to claim 1, characterized in that, Determining that the operation preconditions are maintained and / or utilized by the second motion segment includes: determining that the second motion segment maintains the operation preconditions during execution based on at least one of instrument identification, target identification, temporal or operation sequence association, spatial relationship continuity, posture continuity, and instrument-target interaction continuity, and performing interactions for the training instrument and physical training target associated with the operation preconditions.

4. The method according to claim 1, characterized in that, When the interaction between the device and the target in a motion segment directly causes the training task to transition from the first task state to the second task state, the motion segment is identified as a motion segment that directly contributes to the transition of the corresponding target state.

5. The method according to claim 1, characterized in that, After detecting that the training task has entered an error state, when a motion segment causes the training task to transition from the error state to a preset legal and continueable state, the motion segment is identified as a motion segment that contributes to recovery.

6. The method according to claim 1, characterized in that, A motion segment is defined as a non-contributing motion segment when it fails to directly complete the target state transition, fails to form the operational preconditions that can be actually used by subsequent target state transitions, and fails to restore the training task from an erroneous state to a legal and continueable state.

7. The method according to claim 1, characterized in that, The association between the first motion segment and the second motion segment is determined by a time association window and / or an operation sequence association window, wherein the operation sequence association window allows for at least one intermediate motion segment between the first motion segment and the second motion segment.

8. The method according to claim 1, characterized in that, A target state transition is associated with two or more motion segments, wherein the two or more motion segments include at least one directly contributing motion segment and / or at least one enabling contributing motion segment.

9. The method according to claim 1, characterized in that, The laparoscopic skill evaluation results include at least one of the following: target state transition power, state transition motion efficiency, direct contribution motion volume, enabling contribution motion volume, recovery contribution motion volume, non-contribution motion volume, motion behavior distribution in different state transition stages, and explanatory information corresponding to the motion behavior distribution.

10. The method according to claim 1, characterized in that, The motion data of the training device is obtained through at least one of optical vision, visual marking, RGB-D depth vision, electromagnetic positioning, mechanical coding, inertial sensing, robot encoder, or multi-sensor fusion; the state data of the physical training target is obtained through at least one of RGB image, depth image, visual marking, image target detection, image segmentation, contact sensing, force sensing, position sensing, switch quantity sensing, or multi-source fusion.

11. A laparoscopic state transition contribution recognition and skill evaluation system, characterized in that, include: The motion data acquisition module is used to acquire motion data of the training equipment during laparoscopic physical training. The target object state acquisition module is used to acquire the state data of the entity training target object; The task state transition module is used to determine the first task state, the second task state, and the target state transition according to the task state transition rules. The motion segmentation module is used to divide continuous motion into multiple motion segments; The precondition determination module is used to determine the operation preconditions for the formation of the first motion segment for a first motion segment that does not directly trigger the target state transition; The contribution attribution module is used to determine the subsequent second motion segment, detect whether the operation preconditions are maintained and / or utilized during the execution of the second motion segment and whether the second motion segment actually completes the target state transition, and determine the first motion segment as a motion segment with enabling contribution when the operation preconditions are maintained and / or utilized and the target state transition is actually completed. The skills evaluation module is used to generate laparoscopic skills evaluation results based on the state transition contribution relationship between motion segments and task state transitions.

12. An electronic device comprising a processor and a memory, characterized in that, The memory stores a computer program, which, when executed by the processor, causes the electronic device to perform the method described in any one of claims 1 to 10.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method according to any one of claims 1 to 10.

14. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method according to any one of claims 1 to 10.