Dynamic boundary identification method for surgical sterile area based on fuzzy game theory

Abstract

The invention discloses a dynamic boundary identification method for a surgical sterile area based on a fuzzy game theory, and the method comprises the following steps: S1, collecting spatial position information and a continuous motion track of a participant, and carrying out the preprocessing; s2, initializing a sterile region boundary, and extracting a relative spatial relationship between the participant and the initial sterile region boundary; s3, behavior influence factors are calculated, and weighting processing is carried out; s4, constructing a dynamic fuzzy membership function, and calculating a fuzzy influence value of the participant on the sterile region boundary; s5, introducing an improved evolutionary game mechanism to solve a game equilibrium solution; s6, judging the change trend of the sterile area boundary, and if an updating condition is met, correcting the position of the sterile area boundary; and S7, calculating a border-crossing score value, and if the score value exceeds a threshold value, triggering border-crossing early warning. According to the method, fuzzy logic and an improved evolutionary game mechanism are fused, dynamic boundary recognition and adjustment of the operation sterile area are achieved, and the method has the advantages of being high in intelligence, timely in response and accurate in judgment.

Description

Technical Field

[0001] The present invention relates to the field of intelligent recognition, and in particular to a method for dynamic boundary recognition of a surgical sterile area based on fuzzy game theory. Background Art

[0002] In modern surgical procedures, the management of the surgical sterile area is of great significance for reducing the risk of intraoperative infection and ensuring the safety of patients' lives. Traditional operating room sterility control mainly relies on static division and manual supervision, that is, by setting physical marks or visual warning lines around the operating table in advance to indicate the scope of the sterile area, and requiring medical staff to follow preset paths and operating specifications during the operation to avoid entering or touching the sterile area. However, with the increase in personnel mobility during complex surgical procedures and the popularization of robot-assisted surgical systems, relying solely on static area demarcation and manual experience judgment can no longer meet the needs of accurate monitoring and immediate response to sterile areas in dynamic environments.

[0003] One type of existing research work attempts to improve the degree of automation of sterile area management through computer vision and sensor fusion technology. For example, some systems introduce depth cameras, infrared sensors or RGB-D vision devices to monitor the movement behavior of various subjects in the surgical scene in real time. Through image segmentation and target detection algorithms, the position and posture of medical staff or instruments are automatically identified, and compared with the preset sterile boundary to determine whether there is any cross-border behavior. This type of method has improved the recognition efficiency to a certain extent, but most of them are still based on static boundary models and lack the ability to model the dynamic changes of the sterile area itself. It is especially difficult to deal with the expansion, compression or deformation of the sterile area due to surgical needs. In addition, traditional spatial distance judgment methods mostly use fixed thresholds or binary judgment logic, which cannot express the ambiguity and uncertainty characteristics existing in the actual surgical environment, and are prone to misjudgment or omission.

[0004] Another type of research attempts to introduce behavioral modeling mechanisms to analyze the behavioral trajectories of personnel or equipment to predict whether they are likely to come into contact with sterile areas. However, such methods are often based on probability statistics or simple time series modeling, and it is difficult to comprehensively consider the complex coupling effects between the interactive relationships between subjects, spatial proximity, and behavioral intentions. The behaviors of different subjects in the surgical environment will affect each other. For example, a change in the movement path of a nurse may cause another doctor to have to deviate from the original path and approach the sterile area. If this mutual influence is not modeled, the judgment model will not be able to accurately identify potential risks.

[0005] In addition, existing methods generally lack intelligent feedback and strategy adjustment capabilities. Most systems only output out-of-bounds results or issue warnings in a unidirectional manner, but are unable to dynamically adjust the boundaries of the sterile area based on the current scenario or guide the participating entities to optimize their behavior paths. Especially in intelligent surgical environments that include multiple entities, wearable devices, and robot collaboration, a dynamic recognition mechanism with adaptability, reasoning, and strategy feedback capabilities is needed to replace the traditional static boundary monitoring logic.

[0006] In this context, researchers have attempted to introduce fuzzy logic into the problem of sterile area identification to characterize the uncertainty in spatial approach behavior. For example, the spatial distance between a person and the boundary is mapped to a fuzzy membership degree to determine the possibility of crossing the boundary. However, most existing fuzzy models are single-variable models that do not consider the combined influence of behavioral characteristics (such as length of stay and movement trends), and their modeling capabilities are still relatively limited. Game theory, especially evolutionary game theory, is widely used in multi-agent decision-making behavior modeling in unmanned systems, robot swarms, and traffic path prediction. It can reflect the strategic game process between individuals and its evolutionary results. However, in surgical scenarios, there is still a lack of effective solutions to effectively combine fuzzy logic and evolutionary game mechanisms to achieve intelligent perception and adjustment of the changing trend of sterile area boundaries under the influence of the dynamic behavior of multiple participating subjects.

[0007] In summary, the existing technologies mainly have the following problems: First, the static sterile area boundary model is difficult to adapt to the dynamically changing spatial layout and behavioral distribution during the operation; second, the cross-border identification method based on distance or image processing cannot fully express the ambiguity and uncertainty in the approach behavior of the sterile area; third, there is a lack of comprehensive modeling means that integrates spatial position, behavioral intention and subject interaction relationship, making it difficult to accurately assess the risk of cross-border; fourth, the existing system cannot dynamically correct the sterile boundary based on the subject's behavioral state feedback, and cannot provide behavioral guidance at the strategy evolution level; fifth, there is a lack of a dynamic identification mechanism that can handle the collaborative behavior of multiple participating subjects and integrate fuzzy modeling and evolutionary reasoning.

[0008] Therefore, there is an urgent need for a dynamic identification method that can perform temporal modeling, fuzzy expression and strategy deduction on surgical participants. It can combine the subject's behavior trajectory, spatial proximity and fuzzy risk assessment to infer the current sterile area boundary status in real time, and dynamically issue warnings or correct boundaries when the risk of crossing the boundary increases. Summary of the Invention

[0009] One purpose of the present invention is to propose a method for identifying dynamic boundaries of surgical sterile areas based on fuzzy game theory. The present invention integrates fuzzy logic and an improved evolutionary game mechanism to realize the identification and adaptive adjustment of dynamic boundaries of surgical sterile areas. By modeling and reasoning about the behavior of medical staff and equipment, it can perceive the risk of crossing the boundary in real time and dynamically correct the boundary status. It has the advantages of high judgment accuracy, strong environmental adaptability, and timely risk warning. It can effectively improve the level of spatial safety management in complex surgical scenes, reduce the risk of infection, and enhance the stability and intelligent control capabilities of the system under multi-subject collaborative operation.

[0010] According to an embodiment of the present invention, a method for identifying dynamic boundaries of a surgical sterile area based on fuzzy game theory includes the following steps:

[0011] S1. Collect the spatial position information and continuous motion trajectory of the subjects during the operation, pre-process the collected data, and generate a dynamic behavior data sequence;

[0012] S2. Build a spatial representation model of the surgical scene based on the dynamic behavior data sequence, set the initial sterile zone boundary in the spatial representation model, and extract the relative spatial relationship between the participating subjects and the initial sterile zone boundary at each time step;

[0013] S3. Calculate the behavior influence factor based on the spatial proximity and residence time between the participant and the initial sterile zone boundary, and perform weighted processing on the behavior influence factor to generate a behavior weight;

[0014] S4. Based on the behavior weight and relative spatial relationship, a dynamic fuzzy membership function is constructed to calculate the fuzzy influence value of the participating subjects on the sterile zone boundary;

[0015] S5. Introducing an improved evolutionary game mechanism, using fuzzy influence values as the basis for strategy adjustment, setting the strategy set and immediate payoff function of the participating entities, and solving the game equilibrium solution based on the state evolution process;

[0016] S6. Determine the changing trend of the sterile zone boundary based on the game equilibrium solution. If the update condition is met, dynamically modify the sterile zone boundary position to generate the boundary recognition result at the current moment.

[0017] S7. Use the boundary recognition result as a judgment benchmark to calculate the boundary crossing score value. If the boundary crossing score value of any participating entity exceeds the judgment threshold, the boundary crossing warning mechanism is triggered.

[0018] Optionally, the participating entities include medical staff entities and surgical instrument entities.

[0019] Optionally, the continuous motion trajectory is a sequence of spatial position changes of the participating subject collected continuously in chronological order during the operation.

[0020] Optionally, the preprocessing includes format conversion, noise removal, time alignment and standardization.

[0021] Optionally, the S2 specifically includes:

[0022] S21. Extract the normalized spatial state vector of each participant at each time step from the dynamic behavior data sequence And deconstructed into a three-dimensional space position vector:

[0023] p i (t)=(x i (t),y i (t),z i (t));

[0024] Among them, p i (t) represents the spatial position of participant i at time step t, x i (t), y i (t) and z i (t) represents the coordinate values of participant i in the x, y and z axis directions respectively;

[0025] S22. Combine the three-dimensional spatial position vectors of all participating entities at each time step to form a spatial state frame sequence of the time step:

[0026] P(t)={p1(t),p2(t),...,p N (t)};

[0027] Among them, P(t) represents the spatial state frame sequence, describing the spatial position set of all participating entities at time step t, p N (t) represents the spatial position of the participant N at time step t, and N represents the total number of participants;

[0028] S23. Build a spatial representation model of the surgical scene based on the spatial state frame sequence, set the sterile zone boundary at the initial time step, and define the initial sterile zone boundary as a hexahedron bounding box:

[0029] B0={(x min ,x max ),(y min ,y max ),(z min ,z max )};

[0030] Among them, B0 represents the initial sterile area bounding box, x min and x max Indicates the minimum and maximum values of the sterile area in the x-axis direction, y min and y maxIndicates the minimum and maximum values of the sterile area in the y-axis direction, z min and z max Indicates the minimum and maximum values of the sterile area in the z-axis direction;

[0031] S24. At each time step, calculate the spatial distance between the position of the participating subject and the initial sterile zone boundary box, and define the minimum boundary distance:

[0032]

[0033] Among them, d i (t) represents the minimum boundary distance, which describes the minimum Euclidean distance between the participant i and the initial sterile area boundary box at time step t. represents the boundary point set of the initial sterile area bounding box, b represents the boundary points of the initial sterile area bounding box, and ‖·‖ represents the Euclidean distance.

[0034] Optionally, the S3 specifically includes:

[0035] S31. At each time step, the spatial proximity between the participant and the boundary is calculated based on the spatial position vector of the participant and the initial sterile zone boundary:

[0036] c i (t) = exp(-α·d i (t));

[0037] Among them, c i (t) represents the spatial proximity of participant i at time step t, α represents the distance sensitivity coefficient, d i (t) represents the minimum boundary distance, exp represents the natural exponential function;

[0038] S32. Count the length of time the participant spends near the initial sterile zone boundary in all time steps:

[0039]

[0040] Among them, T i represents the cumulative length of time that participant i stays within the initial sterile area boundary, δ i (t) represents a binary indicator function. When the minimum boundary distance of participant i at time step t is less than the preset threshold δ i (t) = 1, otherwise 0, M represents the total number of time steps;

[0041] S33. Integrate spatial proximity and cumulative dwell time to define the behavior influencing factor:

[0042]

[0043] Among them, fi (t) represents the behavioral influence factor of participant i at time step t, β1 and β2 represent the weighted coefficients of spatial proximity and residence time, respectively, and c i (t) represents the spatial proximity of participant i at time step t, T i Indicates the cumulative length of stay;

[0044] S34. Construct a time decay function to perform weighted summation on the behavior influencing factors to obtain the final behavior weight:

[0045]

[0046] Among them, F i represents the final behavior weight of participant i, f i (t) represents the behavioral influence factor of participant i at time step t, and λ represents the time decay factor.

[0047] Optionally, the S4 specifically includes:

[0048] S41. Combining the behavior weights of the participants and the minimum boundary distance to form a fuzzy input vector:

[0049] X i (t)=(F i ,d i (t));

[0050] Among them, X i (t) represents the fuzzy input vector of participant i at time step t, F i represents the behavior weight of the participating subject, d i (t) represents the minimum boundary distance;

[0051] S42. Define the behavioral risk membership function and the spatial proximity membership function based on the fuzzy input vector and the minimum boundary distance, respectively. Both use the Gaussian membership function form:

[0052]

[0053] Among them, μ A (X i ) represents the behavioral risk membership function, μ B (d i ) represents the spatial proximity membership function, μ X and μ d They represent the mean parameters of behavior weight and boundary distance, σ X and σ d They represent the corresponding standard deviation control parameters, exp represents the natural exponential function;

[0054] S43. Based on fuzzy logic rules, the two membership functions are combined for inference, and the fuzzy influence value is calculated using the product rule:

[0055] μ i (t) = μ A (X i )·μ B (d i );

[0056] Among them, μ i (t) represents the fuzzy influence value of participant i on the sterile zone boundary at time step t;

[0057] S44. Use a smooth output function to perform sliding window smoothing on the fuzzy impact value to alleviate the impact of sudden shift on boundary judgment:

[0058]

[0059] in, represents the smooth fuzzy influence value of participant i at time step t, K represents the number of time steps of the sliding window, μ i (tk) represents the fuzzy influence value of participant i on the sterile zone boundary at time step tk.

[0060] Optionally, the S5 specifically includes:

[0061] S51. The fuzzy influence value is introduced as an input variable into the improved evolutionary game mechanism, and the boundary-related strategy set of the participant i at time step t is set as follows:

[0062] S i ={s1,s2,s3,s4};

[0063] Among them, S i Represents a set of border-related strategies, s1 represents the strategy of maintaining a safe distance, s2 represents the strategy of slowly approaching the border, s3 represents the strategy of staying at the border for a short time, and s4 represents the strategy of crossing the border;

[0064] S52. Initialize the strategy probability distribution of the participating entities:

[0065] π i (t) = {π i,1 (t),π i,2 (t),π i,3 (t),π i,4 (t)};

[0066] Among them, π i (t) represents the strategy probability distribution of participant i at time step t, π i,j (t) represents the strategy s chosen by participant i jThe probability value of , j∈{1,2,3,4}, and satisfies the probability normalization;

[0067] S53. Construct an instant benefit function including fuzzy penalty terms:

[0068]

[0069] in, Indicates that the participant i chooses strategy s at time step t j The instant benefit, r0 represents the basic benefit constant, γ1 represents the fuzzy impact penalty coefficient, represents the fuzzy influence value of participant i at time step t, γ2 represents the strategy cost weight coefficient, C(s j ) represents strategy s j The cost value of

[0070] S54, the strategy cost C (s j ) Assign values according to risk levels:

[0071] C(s1)=0.0, C(s2)=0.2, C(s3)=0.5, C(s4)=1.0;

[0072] Among them, C(s j ) represents strategy s j The cost value of the strategy is as follows: the larger the value, the higher the risk of sterility breach. The risk levels of the four strategies increase in sequence;

[0073] S55. Calculate the expected return of the strategy of participant i at time step t:

[0074]

[0075] in, represents the expected strategic benefit of participant i at time step t, π i,j (t) represents the strategy s chosen by participant i j The probability value of Indicates that the participant i chooses strategy s at time step t j immediate benefits;

[0076] S56. Update the probability distribution of the strategy using a replication dynamic mechanism that introduces fuzzy influence feedback:

[0077]

[0078] Among them, π i,j (t+1) indicates that the participant i chooses strategy s at time step t+1 j The probability value of ,η represents the learning rate of replication dynamics, Indicates that the participant i chooses strategy s at time step tj immediate benefits, represents the expected strategic benefit of participant i at time step t;

[0079] S57. Repeat the strategy evolution process until the strategy probabilities of all participants converge, and record the strategy probability distribution in the converged state as the game equilibrium solution.

[0080] Optionally, the S6 specifically includes:

[0081] S61. Obtaining the game equilibrium solution Calculate the risk probability of crossing the boundary of participant i at time step t:

[0082]

[0083] in, represents the risk probability of crossing the boundary of participant i at time step t, Indicates that after convergence, the participant i chooses strategy s at time step t j The probability value of Indicates a set of cross-border related policies, including s3 and s4;

[0084] S62. Set the cross-border risk judgment threshold θ. If:

[0085]

[0086] Among them, θ represents the threshold for judging the risk of crossing the boundary, If any participant i meets the risk probability of crossing the boundary exceeding the threshold θ, the boundary update mechanism will be triggered;

[0087] S63. Calculate the overall boundary pressure value based on the fuzzy influence values and minimum boundary distances of all participating entities:

[0088]

[0089] Where Ω(t) represents the overall boundary pressure value of all participating entities at time step t, represents the fuzzy influence value of participant i at time step t, d i (t) represents d i (t) represents the minimum boundary distance, α represents the distance sensitivity coefficient, and N represents the total number of participating entities;

[0090] S64. Update the sterile zone boundary position based on the overall boundary pressure value:

[0091]

[0092] Among them, B(t+1) represents the boundary position of the sterile area at time step t+1, B(t) represents the boundary position of the sterile area at time step t+1, Indicates the estimated direction of the gradient of the boundary influence, and λ represents the boundary adjustment step coefficient;

[0093] S65. Use the corrected boundary B(t+1) as the current sterile area boundary recognition result.

[0094] Optionally, the S7 specifically includes:

[0095] S71. Receive the current sterile area boundary recognition result as a judgment basis;

[0096] S72. For each participant, obtain the current fuzzy influence value, and calculate the boundary crossing score value by combining the participant and the minimum boundary distance;

[0097] S73. Set a threshold for determining a cross-border score, and determine whether any participant's cross-border score exceeds the threshold. If so, the participant is deemed to be in a cross-border risk state.

[0098] S74. Generate real-time warning information for the participating entities that meet the cross-border risk conditions, where the warning information includes the participating main number, the cross-border score value, and the location information in the spatial scene.

[0099] The beneficial effects of the present invention are:

[0100] First of all, by introducing fuzzy game theory, the present invention proposes a dynamic boundary recognition method for sterile areas for surgical scenarios. On the basis of solving the problems of static demarcation, fixed rules, and judgment lag in the existing technology, it achieves a more intelligent, adaptive and refined sterile area management effect. Compared with the traditional sterile control method that relies on static area setting and visual judgment, the present invention can comprehensively analyze the spatial position, motion trajectory, behavioral intention and other multi-dimensional information of the participating subjects during the operation, dynamically identify and update the boundary status of the sterile area in real time, and effectively deal with the problem of sterile boundary changes caused by factors such as personnel movement, instrument movement, and space compression in complex scenes.

[0101] Secondly, by constructing behavioral weights and spatial relationships as fuzzy inputs, and then forming a fuzzy membership function and inferring the fuzzy influence value, the present invention effectively improves the model's ability to deal with the behavioral ambiguity and spatial uncertainty of participating entities, and overcomes the roughness and misjudgment risk of traditional binary judgment logic in practical applications. At the same time, the introduction of an improved evolutionary game mechanism enables the system to have strategy evolution and risk adjustment functions, which can not only simulate the behavioral choices and game relationships between participating entities, but also drive the adaptive adjustment of strategy distribution through risk-weighted feedback, thereby realizing feedforward control of out-of-bounds trends.

[0102] Finally, the dynamic boundary adjustment module constructed based on fuzzy influence value and strategy equilibrium solution makes the sterile boundary no longer a static and immutable geometric framework, but a flexible area that responds and adjusts with changes in the subject's behavior state, significantly improving the system's adaptability to nonlinear boundary disturbances in complex surgical environments. Furthermore, the out-of-bounds warning mechanism can perform real-time identification and prompts based on the fuzzy scoring results, effectively ensuring operational safety and collaborative efficiency during the operation.

[0103] In general, the present invention not only realizes the organic integration of fuzzy logic and game theory at the theoretical level, but also provides an integrated sterile area identification solution with dynamic modeling, risk assessment, autonomous adjustment and intelligent feedback capabilities at the application level, providing key technical support for the construction of smart operating rooms and intelligent perception-based medical systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0104] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0105] Figure 1 This is a flow chart of a method for identifying dynamic boundaries of surgical sterile areas based on fuzzy game theory proposed by the present invention;

[0106] Figure 2 This is a diagram of the fuzzy input construction and membership function joint reasoning structure of a dynamic boundary identification method for surgical sterile areas based on fuzzy game theory proposed by the present invention;

[0107] Figure 3 This is a schematic diagram of the strategy evolution and equilibrium solution in the evolutionary game mechanism of the dynamic boundary identification method of the surgical sterile area based on fuzzy game theory proposed by the present invention. DETAILED DESCRIPTION

[0108] The present invention will now be described in further detail with reference to the accompanying drawings, which are simplified schematic diagrams that illustrate the basic structure of the present invention in a schematic manner.

[0109] refer to Figure 1-3 A dynamic boundary recognition method for surgical sterile area based on fuzzy game theory includes the following steps:

[0110] S1. Collect the spatial position information and continuous motion trajectory of the subjects during the operation, pre-process the collected data, and generate a dynamic behavior data sequence;

[0111] S2. Build a spatial representation model of the surgical scene based on the dynamic behavior data sequence, set the initial sterile zone boundary in the spatial representation model, and extract the relative spatial relationship between the participating subjects and the initial sterile zone boundary at each time step;

[0112] S3. Calculate the behavior influence factor based on the spatial proximity and residence time between the participant and the initial sterile zone boundary, and perform weighted processing on the behavior influence factor to generate a behavior weight;

[0113] S4. Based on the behavior weight and relative spatial relationship, a dynamic fuzzy membership function is constructed to calculate the fuzzy influence value of the participating subjects on the sterile zone boundary;

[0114] S5. Introducing an improved evolutionary game mechanism, using fuzzy influence values as the basis for strategy adjustment, setting the strategy set and immediate payoff function of the participating entities, and solving the game equilibrium solution based on the state evolution process;

[0115] S6. Determine the changing trend of the sterile zone boundary based on the game equilibrium solution. If the update condition is met, dynamically modify the sterile zone boundary position to generate the boundary recognition result at the current moment.

[0116] S7. Use the boundary recognition result as a judgment benchmark to calculate the boundary crossing score value. If the boundary crossing score value of any participating entity exceeds the judgment threshold, the boundary crossing warning mechanism is triggered.

[0117] The present invention constructs a dynamic boundary recognition method that integrates behavioral modeling, fuzzy reasoning and game evolution, which can realize continuous perception, intelligent judgment and dynamic correction of the state of the sterile area during surgery, significantly improving the real-time and accuracy of sterile boundary recognition in complex environments, and solving the problems of recognition lag and easy misjudgment in the existing technology.

[0118] In this embodiment, the participating entities include medical staff entities and surgical instrument entities.

[0119] By classifying and modeling medical staff and surgical instruments, the present invention can more specifically identify the spatial behavior patterns of different roles, thereby improving the system's adaptability to the collaborative behavior of multiple types of participants and the refinement of boundary judgment, thereby enhancing the practicality of sterile control.

[0120] In this embodiment, the continuous motion trajectory is a sequence of spatial position changes of the participating subject collected continuously in chronological order during the operation.

[0121] The present invention introduces continuous motion trajectories as a dynamic data source, which helps to comprehensively record the continuous motion changes of participating subjects in the surgical scene, enhance the model's ability to capture behavioral trends, and provide a rich temporal feature basis for subsequent fuzzy modeling and game deduction.

[0122] In this embodiment, the preprocessing includes format conversion, noise removal, time alignment and standardization.

[0123] The present invention improves the quality and timing consistency of behavioral data by setting preprocessing operations such as format conversion, noise removal, time alignment and standardization, provides high-quality input for dynamic identification of sterile areas, and enhances system robustness and calculation accuracy.

[0124] In this embodiment, S2 specifically includes:

[0125] S21. Extract the normalized spatial state vector of each participant at each time step from the dynamic behavior data sequence And deconstructed into a three-dimensional space position vector:

[0126] p i (t)=(x i (t),y i (t),z i (t));

[0127] Among them, p i (t) represents the spatial position of participant i at time step t, x i (t), y i (t) and z i (t) represents the coordinate values of participant i in the x, y and z axis directions respectively;

[0128] S22. Combine the three-dimensional spatial position vectors of all participating entities at each time step to form a spatial state frame sequence of the time step:

[0129] P(t)={p1(t),p2(t),...,p N (t)};

[0130] Among them, P(t) represents the spatial state frame sequence, describing the spatial position set of all participating entities at time step t, p N (t) represents the spatial position of the participant N at time step t, and N represents the total number of participants;

[0131] S23. Build a spatial representation model of the surgical scene based on the spatial state frame sequence, set the sterile zone boundary at the initial time step, and define the initial sterile zone boundary as a hexahedron bounding box:

[0132] B0={(x min ,x max ),(y min ,y max ),(z min ,z max )};

[0133] Among them, B0 represents the initial sterile area bounding box, x min and x max Indicates the minimum and maximum values of the sterile area in the x-axis direction, y min and y max Indicates the minimum and maximum values of the sterile area in the y-axis direction, z min and z max Indicates the minimum and maximum values of the sterile area in the z-axis direction;

[0134] S24. At each time step, calculate the spatial distance between the position of the participating subject and the initial sterile zone boundary box, and define the minimum boundary distance:

[0135]

[0136] Among them, d i (t) represents the minimum boundary distance, which describes the minimum Euclidean distance between the participant i and the initial sterile area boundary box at time step t. represents the boundary point set of the initial sterile area bounding box, b represents the boundary points of the initial sterile area bounding box, and ‖·‖ represents the Euclidean distance.

[0137] The present invention realizes structured modeling of surgical scenes by constructing spatial representation models such as spatial state vectors, state frame sequences and hexahedral bounding boxes, accurately describes the geometric relationship between participating subjects and sterile boundaries, and provides a quantitative basis for dynamic boundary judgment.

[0138] In this embodiment, S3 specifically includes:

[0139] S31. At each time step, the spatial proximity between the participant and the boundary is calculated based on the spatial position vector of the participant and the initial sterile zone boundary:

[0140] c i (t) = exp(-α·d i (t));

[0141] Among them, c i (t) represents the spatial proximity of participant i at time step t, α represents the distance sensitivity coefficient, d i (t) represents the minimum boundary distance, exp represents the natural exponential function;

[0142] S32. Count the length of time the participant spends near the initial sterile zone boundary in all time steps:

[0143]

[0144] Among them, T i represents the cumulative length of time that participant i stays within the initial sterile area boundary, δ i(t) represents a binary indicator function. When the minimum boundary distance of participant i at time step t is less than the preset threshold δ i (t) = 1, otherwise 0, M represents the total number of time steps;

[0145] S33. Integrate spatial proximity and cumulative dwell time to define the behavior influencing factor:

[0146]

[0147] Among them, f i (t) represents the behavioral influence factor of participant i at time step t, β1 and β2 represent the weighted coefficients of spatial proximity and residence time, respectively, and c i (t) represents the spatial proximity of participant i at time step t, T i Indicates the cumulative length of stay;

[0148] S34. Construct a time decay function to perform weighted summation on the behavior influencing factors to obtain the final behavior weight:

[0149]

[0150] Among them, F i represents the final behavior weight of participant i, f i (t) represents the behavioral influence factor of participant i at time step t, and λ represents the time decay factor.

[0151] The present invention combines spatial proximity and residence time to calculate the behavior influence factor, and introduces a time decay function to form the behavior weight, which can more realistically express the potential impact of the participating subjects on the sterile area and enhance the sensitivity of boundary judgment to changes in behavioral characteristics.

[0152] In this embodiment, the S4 specifically includes:

[0153] S41. Combining the behavior weights of the participants and the minimum boundary distance to form a fuzzy input vector:

[0154] X i (t)=(F i ,d i (t));

[0155] Among them, X i (t) represents the fuzzy input vector of participant i at time step t, F i represents the behavior weight of the participating subject, d i (t) represents the minimum boundary distance;

[0156] S42. Define the behavioral risk membership function and the spatial proximity membership function based on the fuzzy input vector and the minimum boundary distance, respectively. Both use the Gaussian membership function form:

[0157]

[0158] Among them, μ A (X i ) represents the behavioral risk membership function, μ B (d i ) represents the spatial proximity membership function, μ X and μ d They represent the mean parameters of behavior weight and boundary distance, σ X and σ d They represent the corresponding standard deviation control parameters, exp represents the natural exponential function;

[0159] S43. Based on fuzzy logic rules, the two membership functions are combined for inference, and the fuzzy influence value is calculated using the product rule:

[0160] μ i (t) = μ A (X i )·μ B (d i );

[0161] Among them, μ i (t) represents the fuzzy influence value of participant i on the sterile zone boundary at time step t;

[0162] S44. Use a smooth output function to perform sliding window smoothing on the fuzzy impact value to alleviate the impact of sudden shift on boundary judgment:

[0163]

[0164] in, represents the smooth fuzzy influence value of participant i at time step t, K represents the number of time steps of the sliding window, μ i (tk) represents the fuzzy influence value of participant i on the sterile zone boundary at time step tk.

[0165] The present invention constructs a Gaussian membership function through fuzzy input, integrates behavioral weight and spatial distance for fuzzy reasoning, and introduces a sliding smoothing mechanism, which significantly improves the system's fault tolerance, stability and fuzzy boundary expression ability in complex behavioral environments.

[0166] In this embodiment, the S5 specifically includes:

[0167] S51. The fuzzy influence value is introduced as an input variable into the improved evolutionary game mechanism, and the boundary-related strategy set of the participant i at time step t is set as follows:

[0168] S i ={s1,s2,s3,s4};

[0169] Among them, S i Represents a set of border-related strategies, s1 represents the strategy of maintaining a safe distance, s2 represents the strategy of slowly approaching the border, s3 represents the strategy of staying at the border for a short time, and s4 represents the strategy of crossing the border;

[0170] S52. Initialize the strategy probability distribution of the participating entities:

[0171] π i (t) = {π i,1 (t),π i,2 (t),π i,3 (t),π i,4 (t)};

[0172] Among them, π i (t) represents the strategy probability distribution of participant i at time step t, π i,j (t) represents the strategy s chosen by participant i j The probability value of , j∈{1,2,3,4}, and satisfies the probability normalization;

[0173] S53. Construct an instant benefit function including fuzzy penalty terms:

[0174]

[0175] in, Indicates that the participant i chooses strategy s at time step t j The instant benefit, r0 represents the basic benefit constant, γ1 represents the fuzzy impact penalty coefficient, represents the fuzzy influence value of participant i at time step t, γ2 represents the strategy cost weight coefficient, C(s j ) represents strategy s j The cost value of

[0176] S54, the strategy cost C (s j ) Assign values according to risk levels:

[0177] C(s1)=0.0, C(s2)=0.2, C(s3)=0.5, C(s4)=1.0;

[0178] Among them, C(s j ) represents strategy s jThe cost value of the strategy is as follows: the larger the value, the higher the risk of sterility breach. The risk levels of the four strategies increase in sequence;

[0179] S55. Calculate the expected return of the strategy of participant i at time step t:

[0180]

[0181] in, represents the expected strategic benefit of participant i at time step t, π i,j (t) represents the strategy s chosen by participant i j The probability value of Indicates that the participant i chooses strategy s at time step t j immediate benefits;

[0182] S56. Update the probability distribution of the strategy using a replication dynamic mechanism that introduces fuzzy influence feedback:

[0183]

[0184] Among them, π i,j (t+1) indicates that the participant i chooses strategy s at time step t+1 j The probability value of ,η represents the learning rate of replication dynamics, Indicates that the participant i chooses strategy s at time step t j immediate benefits, represents the expected strategic benefit of participant i at time step t;

[0185] S57. Repeat the strategy evolution process until the strategy probabilities of all participants converge, and record the strategy probability distribution in the converged state as the game equilibrium solution.

[0186] The present invention introduces fuzzy influence value participation strategy modeling, and combines the strategy cost value and evolution mechanism of risk level setting to dynamically adjust the behavior strategy distribution of participating entities, thereby improving the system's response capability to boundary risk changes and decision rationality.

[0187] In this embodiment, S6 specifically includes:

[0188] S61. Obtaining the game equilibrium solution Calculate the risk probability of crossing the boundary of participant i at time step t:

[0189]

[0190] in, represents the risk probability of crossing the boundary of participant i at time step t, Indicates that after convergence, the participant i chooses strategy s at time step t j The probability value of Indicates a set of cross-border related policies, including s3 and s4;

[0191] S62. Set the cross-border risk judgment threshold θ. If:

[0192]

[0193] Among them, θ represents the threshold for judging the risk of crossing the boundary, If any participant i meets the risk probability of crossing the boundary exceeding the threshold θ, the boundary update mechanism will be triggered;

[0194] S63. Calculate the overall boundary pressure value based on the fuzzy influence values and minimum boundary distances of all participating entities:

[0195]

[0196] Where Ω(t) represents the overall boundary pressure value of all participating entities at time step t, represents the fuzzy influence value of participant i at time step t, d i (t) represents d i (t) represents the minimum boundary distance, α represents the distance sensitivity coefficient, and N represents the total number of participating entities;

[0197] S64. Update the sterile zone boundary position based on the overall boundary pressure value:

[0198]

[0199] Among them, B(t+1) represents the boundary position of the sterile area at time step t+1, B(t) represents the boundary position of the sterile area at time step t+1, Indicates the estimated direction of the gradient of the boundary influence, and λ represents the boundary adjustment step coefficient;

[0200] S65. Use the corrected boundary B(t+1) as the current sterile area boundary recognition result.

[0201] The present invention dynamically adjusts the sterile area boundary based on the game equilibrium result and the boundary pressure value, realizes the adaptive evolution control mechanism of the boundary state, avoids the inflexibility brought by static rules, and effectively improves the level of space safety management during the operation.

[0202] In this embodiment, the S7 specifically includes:

[0203] S71. Receive the current sterile area boundary recognition result as a judgment basis;

[0204] S72. For each participant, obtain the current fuzzy influence value, and calculate the boundary crossing score value by combining the participant and the minimum boundary distance;

[0205] S73. Set a threshold for determining a cross-border score, and determine whether any participant's cross-border score exceeds the threshold. If so, the participant is deemed to be in a cross-border risk state.

[0206] S74. Generate real-time warning information for the participating entities that meet the cross-border risk conditions, where the warning information includes the participating main number, the cross-border score value, and the location information in the spatial scene.

[0207] The present invention realizes a closed-loop response from out-of-bounds recognition to real-time prompts through a fuzzy scoring-driven out-of-bounds warning mechanism, ensuring surgical safety and providing timely feedback to medical staff, thereby improving the system's human-computer interaction efficiency and practical application value.

[0208] Example 1:

[0209] To verify the feasibility of this invention, a month-long field test and comparative experiment was conducted in a smart operating room environment at a Class A tertiary hospital. The operating room was equipped with a three-dimensional spatial camera system, medical-grade inertial sensors, robotic arm assistive equipment, and a real-time data acquisition terminal. The experiment focused on three types of surgeries with high dynamic participation: thoracoscopic-assisted heart valve repair, cerebral aneurysm resection, and complex laparoscopic gastrointestinal reconstruction. The experiment evaluated the effectiveness of the proposed "dynamic boundary identification method for surgical sterile areas based on fuzzy game theory" in a complex and practical environment.

[0210] In the above-mentioned test scenario, traditional sterile area management relies on statically set boundary boxes before surgery, ground stickers, and manual monitoring during surgery, and lacks the ability to respond to changes in personnel and equipment behavior. Past experience has shown that the problem of blurred sterile boundaries frequently occurs during surgery due to reasons such as patient position adjustment and equipment rearrangement by medical staff. Especially in the process of high-density collaborative operations, the situation where surgical instrument arms or nurses mistakenly enter the sterile area becomes a potential risk point.

[0211] To address this issue, a complete dynamic recognition system was deployed in the operating room. The system consists of three components: a front-end data acquisition module, a mid-end boundary modeling and recognition reasoning module, and a back-end intelligent early warning module. During surgery, the system collects the three-dimensional positions of medical staff and surgical instruments every 0.5 seconds. By integrating inertial navigation information, it constructs continuous motion trajectories, forming a dynamic behavior data sequence. Subsequently, during the modeling phase, the system sets the initial sterile boundary position and, based on the spatial proximity and duration of the participants, calculates behavioral influence factors and generates behavioral weights. The behavioral weights and spatial distances are then input into a fuzzy logic model to generate fuzzy influence values.

[0212] In actual operation, fuzzy impact values drive the evolution of the strategic game process, the behavioral strategies of the participants are adjusted according to the risk level, and the game equilibrium results reflect the trend of boundary changes. The system updates the sterile zone boundaries in real time accordingly. Finally, the out-of-bounds scoring mechanism combines the game output and the fuzzy value to issue early warning prompts for abnormal behavior. All warning events, out-of-bounds scores and identification data are recorded for subsequent analysis.

[0213] In the experimental setup, 16 real surgeries were selected for testing, of which 8 used the system of the present invention for dynamic boundary recognition, and the other 8 still used manual supervision and static sterile area setting as the control group. By deploying independent observers during the operation and using postoperative video backtracking to record and count the key indicators, the number of boundary violations, warning response time, actual sterile area destruction events, and surgical interference were analyzed.

[0214] Table 1 Performance comparison table

[0215]

[0216] In terms of boundary update response, the system of the present invention achieves a dynamic boundary update every 6.1 seconds on average. Compared with the average 63.4 seconds required by traditional methods that rely on manual judgment and static settings, the overall response speed is increased by nearly ten times. This high-frequency, high-real-time boundary adjustment mechanism significantly enhances the system's adaptability to environmental changes, and effectively addresses the problem of sterile area morphology fluctuations caused by personnel movement and equipment movement during surgery.

[0217] In terms of cross-border risk identification, the present invention uses fuzzy logic modeling and evolutionary game reasoning to successfully warn of 92.7% of cross-border events, while the traditional control group only has 54.2%, demonstrating that the present invention has significant accuracy and robustness in identifying cross-border behaviors. In addition, the system of the present invention only generates an average of 1.2 false positives per surgery, compared to 3.4 false positives under traditional methods, significantly reducing the false alarm rate and avoiding psychological pressure and operation interruptions caused by system false alarms for intraoperative personnel.

[0218] In terms of intraoperative safety, experimental data show that the group of the present invention achieved a control effect of zero actual cross-border destruction events in 16 operations, while the traditional method still had 6 actual cross-border situations, which directly indicates that the present invention has a stronger ability to ensure the integrity of the sterile area. At the same time, there was only 1 intraoperative operation interruption event in the group of the present invention and 4 in the control group, indicating that the system has advantages in ensuring the continuity and rhythm of the operation, and did not interfere with the surgeon. Instead, it optimized the behavioral path through visual guidance.

[0219] More importantly, in terms of postoperative sterility monitoring, no positive contamination cultures were found in the group of the present invention, while 3 contamination prompts were found in the control group, further confirming the effectiveness of the system and its infection prevention and control capabilities. In terms of user experience, the medical staff participating in the experiment gave a subjective satisfaction score of 9.1 points (out of 10 points) for the system of the present invention, which is much higher than the 7.3 points of the traditional solution, indicating that the system has high acceptance and friendly interaction in actual operation, and has good prospects for clinical promotion.

[0220] To sum up, the experimental data fully verifies that the present invention achieves the goals of intelligent, dynamic and high-precision sterile boundary recognition in surgical scenarios, which not only improves the safety of the surgical space, but also enhances the efficiency of human-machine collaboration, reflecting the significant technical value and application advantages of the present invention in intelligent medical scenarios.

[0221] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A method for dynamic boundary recognition of surgical sterile area based on fuzzy game theory, characterized in that: The steps include: S1. Collect the spatial position information and continuous motion trajectory of the subjects during the operation, pre-process the collected data, and generate a dynamic behavior data sequence; S2. Build a spatial representation model of the surgical scene based on the dynamic behavior data sequence, set the initial sterile zone boundary in the spatial representation model, and extract the relative spatial relationship between the participating subjects and the initial sterile zone boundary at each time step; S3. Calculate the behavior influence factor based on the spatial proximity and residence time between the participant and the initial sterile zone boundary, and perform weighted processing on the behavior influence factor to generate a behavior weight; S4. Based on the behavior weight and relative spatial relationship, a dynamic fuzzy membership function is constructed to calculate the fuzzy influence value of the participating subjects on the sterile zone boundary; S5. Introducing an improved evolutionary game mechanism, using fuzzy influence values as the basis for strategy adjustment, setting the strategy set and immediate payoff function of the participating entities, and solving the game equilibrium solution based on the state evolution process; S6. Determine the changing trend of the sterile zone boundary based on the game equilibrium solution. If the update condition is met, dynamically modify the sterile zone boundary position to generate the boundary recognition result at the current moment. S7. Use the boundary recognition result as a judgment benchmark to calculate the boundary crossing score value. If the boundary crossing score value of any participating entity exceeds the judgment threshold, the boundary crossing warning mechanism is triggered.

2. The method for dynamic boundary recognition of surgical sterile area based on fuzzy game theory according to claim 1, characterized in that: The participating entities include medical staff and surgical instruments.

3. The method for dynamic boundary recognition of surgical sterile area based on fuzzy game theory according to claim 1, characterized in that: The continuous motion trajectory is a sequence of spatial position changes of the participating subject collected continuously in chronological order during the operation.

4. The method for dynamic boundary recognition of surgical sterile area based on fuzzy game theory according to claim 1, characterized in that: The preprocessing includes format conversion, noise removal, time alignment and normalization.

5. The method for dynamic boundary recognition of surgical sterile area based on fuzzy game theory according to claim 1, characterized in that: The S2 specifically includes: S21. extracting the normalized spatial state vector of each participant at each time step from the dynamic behavior data sequence, and deconstructing it into a three-dimensional spatial position vector; S22, combining the three-dimensional spatial position vectors of all participating entities at each time step to form a spatial state frame sequence of the time step; S23. constructing a spatial representation model of the surgical scene based on the spatial state frame sequence, setting a sterile area boundary at an initial time step, and defining the initial sterile area boundary as a hexahedral bounding box; S24. At each time step, calculate the spatial distance between the position of the participating subject and the initial sterile area boundary box, and define the minimum boundary distance.

6. The method for dynamic boundary recognition of surgical sterile area based on fuzzy game theory according to claim 1, characterized in that: The S3 specifically includes: S31. At each time step, the spatial proximity between the participant and the boundary is calculated based on the spatial position vector of the participant and the initial sterile zone boundary: c i (t)=exp(-α·d i (t)); Among them, c i (t) represents the spatial proximity of participant i at time step t, α represents the distance sensitivity coefficient, d i (t) represents the minimum boundary distance, exp represents the natural exponential function; S32, counting the length of time the participant spends near the boundary of the initial sterile zone in all time steps; S33. Integrate spatial proximity and cumulative dwell time to define the behavior influencing factor: Among them, f i (t) represents the behavioral influence factor of participant i at time step t, β1 and β2 represent the weighted coefficients of spatial proximity and residence time, respectively, and c i (t) represents the spatial proximity of participant i at time step t, T i Indicates the cumulative length of stay; S34. Construct a time decay function to perform weighted summation on the behavior influencing factors to obtain the final behavior weight.

7. The method for dynamic boundary recognition of surgical sterile area based on fuzzy game theory according to claim 1, characterized in that: The S4 specifically includes: S41, combining the behavior weights of the participating entities and the minimum boundary distance to form a fuzzy input vector; S42. Define the behavioral risk membership function and the spatial proximity membership function based on the fuzzy input vector and the minimum boundary distance, respectively, both of which adopt the form of Gaussian membership function; S43, based on fuzzy logic rules, the two membership functions are jointly reasoned and the fuzzy influence value is calculated using a product rule; S44. Perform sliding window smoothing on the fuzzy impact value using a smoothing output function to alleviate the impact of sudden shift on boundary judgment.

8. The method for dynamic boundary recognition of surgical sterile area based on fuzzy game theory according to claim 1, characterized in that: The S5 specifically includes: S51. The fuzzy influence value is introduced as an input variable into the improved evolutionary game mechanism, and the boundary-related strategy set of the participant i at time step t is set; S52, initializing the strategy probability distribution of the participating entities; S53. Construct an instant benefit function including fuzzy penalty terms: in, Indicates that the participant i chooses strategy s at time step t j The instant benefit, r0 represents the basic benefit constant, γ1 represents the fuzzy impact penalty coefficient, represents the fuzzy influence value of participant i at time step t, γ2 represents the strategy cost weight coefficient, C(s j ) represents strategy s j The cost value of S54, the strategy cost C (s j ) Assign values according to risk levels; S55. Calculate the expected return of the strategy of participant i at time step t: in, represents the expected strategic benefit of participant i at time step t, π i,j (t) represents the strategy s chosen by participant i j The probability value of Indicates that the participant i chooses strategy s at time step t j immediate benefits; S56. Update the probability distribution of the strategy using a replication dynamic mechanism that introduces fuzzy influence feedback: Among them, π i,j (t+1) indicates that the participant i chooses strategy s at time step t+1 j The probability value of ,η represents the learning rate of replication dynamics, Indicates that the participant i chooses strategy s at time step t j immediate benefits, represents the expected strategic benefit of participant i at time step t; S57. Repeat the strategy evolution process until the strategy probabilities of all participants converge, and record the strategy probability distribution in the converged state as the game equilibrium solution.

9. The method for dynamic boundary recognition of surgical sterile area based on fuzzy game theory according to claim 1, characterized in that: The S6 specifically includes: S61. Obtaining the game equilibrium solution Calculate the risk probability of crossing the boundary of participant i at time step t; S62. Set the cross-border risk judgment threshold θ. If: Among them, θ represents the threshold for judging the risk of crossing the boundary, If any participant i meets the risk probability of crossing the boundary exceeding the threshold θ, the boundary update mechanism will be triggered; S63. Calculate the overall boundary pressure value based on the fuzzy influence values and minimum boundary distances of all participating entities: Where Ω(t) represents the overall boundary pressure value of all participating entities at time step t, represents the fuzzy influence value of participant i at time step t, d i (t) represents d i (t) represents the minimum boundary distance, α represents the distance sensitivity coefficient, and N represents the total number of participating entities; S64. Update the sterile zone boundary position based on the overall boundary pressure value; S65. Use the corrected boundary B(t+1) as the current sterile area boundary recognition result.

10. The method for dynamic boundary recognition of surgical sterile area based on fuzzy game theory according to claim 1, characterized in that: The S7 specifically includes: S71. Receive the current sterile area boundary recognition result as a judgment basis; S72. For each participant, obtain the current fuzzy influence value, and calculate the boundary crossing score value by combining the participant and the minimum boundary distance; S73. Set a threshold for determining a cross-border score, and determine whether any participant's cross-border score exceeds the threshold. If so, the participant is deemed to be in a cross-border risk state. S74. Generate real-time warning information for the participating entities that meet the cross-border risk conditions, where the warning information includes the participating main number, the cross-border score value, and the location information in the spatial scene.

Images