A waveform superposition optimization method and system for energy surgical equipment

Abstract

The present invention discloses a waveform superposition optimization method and system for energy surgical equipment, and relates to the field of medical device technology. A waveform superposition optimization system for energy surgical equipment includes: an energy surgical equipment surgical area identification module, an energy surgical equipment waveform superposition module, and an energy surgical equipment smooth wave module. The present invention realizes accurate extraction of surgical area features through the adaptive convolution layer and self-attention mechanism in the surgical area image analysis model, can effectively distinguish healthy tissue, cutting area and coagulation area, and improve the accuracy of surgical area status signal; based on the surgical area status signal and the waveform superposition optimization model, dynamically adjusts the use waveform parameters of the energy surgical equipment; enables the energy surgical equipment to adapt to different surgical area states and reduce unnecessary tissue damage; optimizes the waveform superposition process, eliminates energy fluctuations that may be caused by waveform mutations, optimizes the comfort of waveform changes, and reduces the risk of tissue carbonization or excessive cutting.

Description

Technical Field

[0001] The present invention relates to the technical field of medical devices, and in particular to a waveform superposition optimization method and system for energy surgical equipment. Background Art

[0002] Energy surgical devices achieve precise cutting and rapid coagulation through the thermal effect of high-frequency electric current, and are widely used in surgical and minimally invasive surgeries to improve surgical efficiency, reduce bleeding and reduce the risk of infection. During the cutting and coagulation process, the wound state will change dynamically as the operation progresses, such as tissue drying, carbonization or fluid secretion. These changes will significantly affect the conductivity of the tissue, thereby changing the energy distribution of the high-frequency current. If the energy surgical device maintains the same frequency waveform for output, it may not be able to adapt to the dynamic changes in tissue conductivity, resulting in local heat accumulation, increasing the risk of tissue carbonization, and even damage to surrounding healthy tissue.

[0003] Therefore, in order to meet the medical needs under different wound conditions, it is necessary to accurately identify the current wound state and optimize the high-frequency current waveform amplitude of the energy surgical equipment based on the current wound state. Through the waveform superposition method, the high-frequency waveform and the low-frequency waveform can be dynamically combined to adjust the energy distribution to more accurately meet the requirements of cutting and coagulation. Summary of the invention

[0004] The present invention aims to provide a method and system for optimizing waveform superposition of energy surgical equipment, which can more accurately meet the requirements of cutting and coagulation by dynamically adjusting the output of the energy surgical equipment.

[0005] A method for optimizing waveform superposition of an energy surgical device comprises the following steps:

[0006] Receiving an image of the surgical area to be analyzed corresponding to the energy surgical device; inputting the image of the surgical area to be analyzed into an image analysis model of the surgical area of ​​the energy surgical device for analysis to obtain a surgical area status signal;

[0007] The energy surgery equipment surgical area image analysis model includes an edge cutting layer, an image feature recognition layer and a surgical area state output layer, wherein the image feature recognition layer is used to perform feature recognition on the surgical area image to be analyzed based on the adaptively changed convolution layer and the self-attention mechanism;

[0008] Based on the analysis of the surgical area status signal and the energy surgical equipment waveform superposition optimization model, the energy surgical equipment waveform amplitude signal is obtained; the waveform parameters of the energy surgical equipment are changed according to the waveform amplitude signal of the energy surgical equipment; in the energy surgical equipment waveform superposition optimization model, a signal feature analysis layer, a smooth waveform optimization layer and a waveform superposition output layer are included; wherein the smooth waveform optimization layer is used to smoothly optimize the frequency waveform set to be superimposed output from the signal feature analysis layer, and smoothly connect it with the original energy surgical equipment waveform amplitude signal, so as to optimize the comfort of the energy surgical equipment waveform change; the original energy surgical equipment waveform amplitude signal represents the corresponding waveform parameters of the current energy surgical equipment.

[0009] As a preferred technical solution of the present invention, the energy surgery equipment surgical area image analysis model includes an edge cutting layer, an image feature recognition layer and a surgical area state output layer;

[0010] The edge cutting layer is used to perform edge cutting on the image of the surgical area to be analyzed, so as to obtain a pre-processed image of the surgical area to be analyzed;

[0011] The image feature recognition layer is used to perform feature recognition on the pre-processed surgical area image to be analyzed to obtain the surgical area status signal;

[0012] The image feature recognition layer is built based on the U-net architecture and transformer model;

[0013] The surgical area status output layer is used to output the surgical area status signal.

[0014] As a preferred technical solution of the present invention, the image feature recognition layer includes a complexity calculation layer, an encoder layer, a decoder layer, a skip connection layer and a label classification layer;

[0015] Among them, the jump connection layer contains an attention gate;

[0016] In the complexity calculation layer, it is used to evaluate the complexity of the pre-processed surgical area image to be analyzed and obtain the image complexity score;

[0017] In the encoder layer, M residual blocks are included, each of which is composed of N convolutional layers and 1 residual connection; it is used to extract features of the pre-processed image of the surgical area to be analyzed, and obtain a feature image of the pre-processed surgical area to be analyzed;

[0018] Using the formula The specific value of M is calculated, F represents the image complexity score, and M max Indicates the maximum value of the preset M, M min Indicates the minimum value of the preset M, Indicates floor rounding operation;

[0019] Using the formula The specific value of N is calculated, F represents the image complexity score, N max Indicates the maximum value of preset N, N min Indicates the minimum value of preset N;

[0020] In the skip connection layer, it is used to assign weights to the pre-processed feature image of the technical area to be analyzed, so as to obtain a feature weight image of the pre-processed technical area to be analyzed;

[0021] In the decoder layer, it is used to perform feature fusion on the pre-processed feature weight image of the surgical area to be analyzed to obtain the feature image of the surgical area to be identified;

[0022] In the label classification layer, it is used to perform feature recognition on the feature image of the surgical area to be identified and obtain the surgical area status signal.

[0023] As a preferred technical solution of the present invention, the specific operation of training the image feature recognition layer includes:

[0024] Collecting several groups of surgical area image state recognition training samples; each group of surgical area image state recognition training samples contains a verified state signal and a surgical area image; combining several groups of surgical area image state recognition training samples to obtain a surgical area image state recognition training set;

[0025] The surgical area image state recognition training set is input into the image feature recognition layer for model training to obtain the initial image feature recognition layer; the initial image feature recognition layer is subjected to model evaluation. If the initial image feature recognition layer passes the model evaluation, the initial image feature recognition layer is used as the image feature recognition layer in the surgical area image analysis model of the energy surgical device; otherwise, the surgical area image state recognition training set is used to continue model training.

[0026] As a preferred technical solution of the present invention, the waveform superposition optimization model of the energy surgical device includes a signal feature analysis layer, a smooth waveform optimization layer and a waveform superposition output layer;

[0027] The signal feature analysis layer is used to perform waveform matching on the state signal of the operation area to obtain the frequency waveform set X to be superimposed; where X={X i |i=1,2,…,I},X i represents the frequency waveform to be superimposed in the frequency waveform set X to be superimposed;

[0028] The signal feature analysis layer is constructed based on the SVM model;

[0029] The smooth waveform optimization layer is used to output the signal of the superimposed frequency waveform set X to obtain the waveform amplitude signal used by the energy surgical device;

[0030] The waveform superposition output layer is used to output the waveform amplitude signal used by the energy surgical device.

[0031] As a preferred technical solution of the present invention, the specific steps of outputting signals in the smooth waveform optimization layer include:

[0032] Obtain the waveform amplitude signal used by the original energy surgical device; construct K smoothed wave signal individuals G based on the waveform amplitude signal used by the original energy surgical device and the frequency waveform set X to be superimposed k , k=1,2,…,K; smooth variable wave signal individual G k represents a variable wave superposition signal for energy surgical equipment; K smooth variable wave signal individuals G k Combine to obtain the individual iterative population of the smooth variable wave signal; set the maximum number of iterations;

[0033] Construct a simulated signal wave variation evaluation model to extract the characteristics of the smooth wave variation signal and obtain the characteristics of the smooth wave variation signal; evaluate the comfort level of the smooth wave variation signal and obtain the comfort level of the simulated smooth wave variation signal; use the comfort level of the simulated smooth wave variation signal as the individual smooth wave variation signal G k The fitness S k ;

[0034] The population iteration process is divided into the first population iteration stage and the second population iteration stage;

[0035] In the first stage of population iteration, the smooth wave signal individual with the largest current fitness is selected as the elite smooth wave signal individual in each iteration. When the population is updated, the population is iteratively updated according to the elite smooth wave signal individuals to obtain a new smooth wave signal individual iterative population;

[0036] In the second stage of population iteration, a population mutation factor is introduced; in each iteration, the smooth wave signal individual iteration population is mutated and updated based on the population mutation factor to obtain a new smooth wave signal individual iteration population; the population mutation factor is set based on the Cauchy mutation strategy;

[0037] When the maximum number of iterations is reached, the smooth wave signal individual corresponding to the maximum fitness is output, which is the optimal smooth wave signal individual; based on the optimal smooth wave signal individual, the waveform amplitude signal used by the energy surgical device is output.

[0038] As a preferred technical solution of the present invention, the specific steps of training the signal feature analysis layer include:

[0039] Collecting several groups of signal waveform matching training samples; each group of signal waveform matching training samples contains state signal features and corresponding frequency waveforms; combining several groups of signal waveform matching training samples to obtain a signal waveform matching training set;

[0040] The signal waveform matching training set is input into the SVM model for model training to obtain the initial signal feature analysis layer; the initial signal feature analysis layer is evaluated, and if the initial signal feature analysis layer passes the model evaluation, the initial signal feature analysis layer is used as the signal feature analysis layer in the waveform superposition optimization model of the energy surgical device; otherwise, the signal waveform matching training set is used to continue model training.

[0041] A waveform superposition optimization system for energy surgical equipment, comprising:

[0042] The energy surgery equipment surgical area recognition module includes an image acquisition unit and a feature analysis unit; the image acquisition unit is used to receive the surgical area image to be analyzed corresponding to the energy surgery equipment; the feature analysis unit is used to input the surgical area image to be analyzed into the energy surgery equipment surgical area image analysis model for analysis to obtain the surgical area state signal; the energy surgery equipment surgical area image analysis model includes an edge cutting layer, an image feature recognition layer and a surgical area state output layer, wherein the image feature recognition layer is used to perform feature recognition on the surgical area image to be analyzed based on the adaptively changed convolution layer and the self-attention mechanism;

[0043] The energy surgical equipment waveform superposition module includes a waveform analysis unit; the waveform analysis unit is used to analyze based on the surgical area state signal and the energy surgical equipment waveform superposition optimization model to obtain the energy surgical equipment use waveform amplitude signal; according to the energy surgical equipment use waveform amplitude signal, the use waveform parameters of the energy surgical equipment are changed; in the energy surgical equipment waveform superposition optimization model, a signal feature analysis layer, a smooth waveform optimization layer and a waveform superposition output layer are included;

[0044] The energy surgical equipment smooth wave changing module includes a superposition optimization unit; the superposition optimization unit is used to construct a smooth waveform optimization layer, the smooth waveform optimization layer is used to smoothly optimize the frequency waveform set to be superimposed output in the signal feature analysis layer, and smoothly connect it with the waveform amplitude signal used by the original energy surgical equipment to optimize the comfort of the waveform change of the energy surgical equipment; the waveform amplitude signal used by the original energy surgical equipment represents the corresponding waveform parameters used by the current energy surgical equipment.

[0045] The present invention has the following advantages:

[0046] 1. The present invention realizes accurate extraction of surgical area features through the adaptive convolution layer and self-attention mechanism in the surgical area image analysis model, can effectively distinguish healthy tissue, cutting area and coagulation area, and improve the accuracy of surgical area status signal; dynamically adjusts the waveform parameters of energy surgical equipment based on surgical area status signal and waveform superposition optimization model; enables energy surgical equipment to adapt to different surgical area states, improves cutting and coagulation efficiency, and reduces unnecessary tissue damage; optimizes the waveform superposition process, eliminates energy fluctuations that may be caused by waveform mutations, optimizes the "comfort" of waveform changes, reduces local heat accumulation, and reduces the risk of tissue carbonization or excessive cutting.

[0047] 2. The present invention realizes efficient processing of surgical area images and state signal output by utilizing the image feature recognition layer in combination with complexity assessment, adaptive parameter adjustment, residual block feature extraction and attention mechanism. It has high flexibility, high efficiency and high accuracy. Its design can dynamically adapt to surgical area images of different complexities, reduce computational overhead, improve real-time and robustness, and provide strong technical support for intelligent control of energy surgical equipment and surgical assistance. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A schematic diagram of the structure of a waveform superposition optimization system for energy surgical equipment used in an embodiment of the present invention. DETAILED DESCRIPTION

[0049] In order to enable persons skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0050] Embodiment 1, a method for optimizing waveform superposition of an energy surgical device, comprising the following steps:

[0051] Receive the surgical area image to be analyzed corresponding to the energy surgical device; input the surgical area image to be analyzed into the surgical area image analysis model of the energy surgical device for analysis to obtain the surgical area status signal; the surgical area image to be analyzed refers to the image data of the area of ​​action of the energy surgical device, reflecting the tissue status and dynamic changes of the surgical area, usually including morphology, temperature or other key information. Such images are an important basis for surgical monitoring and status analysis, and provide support for doctors' real-time decision-making and surgical optimization;

[0052] The energy surgery equipment surgical area image analysis model includes an edge cutting layer, an image feature recognition layer and a surgical area state output layer, wherein the image feature recognition layer is used to perform feature recognition on the surgical area image to be analyzed based on the adaptively changed convolution layer and the self-attention mechanism;

[0053] The edge cutting layer is used to perform edge cutting on the surgical area image to be analyzed to obtain a preprocessed surgical area image to be analyzed; the edge cutting layer preprocesses the edge information of the surgical area image, highlights the key area and reduces the interference of background noise, so that the subsequent feature extraction is more focused;

[0054] The image feature recognition layer is used to perform feature recognition on the pre-processed surgical area image to be analyzed to obtain the surgical area status signal;

[0055] The image feature recognition layer is built based on the U-net architecture and the transformer model. Based on the multi-scale feature extraction capability of U-Net, the global and local information are fully utilized to accurately capture the image features of the surgical area. The Transformer module is introduced to enhance the model's modeling capabilities for long-distance dependencies and global context information through the global self-attention mechanism, which is particularly suitable for complex surgical area states.

[0056] The surgical area status output layer is used to output surgical area status signals;

[0057] By combining advanced technologies such as edge cutting layer, U-Net and Transformer, the efficiency, accuracy and robustness of surgical area image feature extraction and state signal output are significantly improved. The model design provides solid technical support for the intelligent control of energy surgical equipment and real-time feedback during surgery, while optimizing the surgical process and improving surgical safety and effectiveness.

[0058] The image feature recognition layer includes a complexity calculation layer, an encoder layer, a decoder layer, a skip connection layer, and a label classification layer;

[0059] Among them, the jump connection layer contains an attention gate;

[0060] In the complexity calculation layer, it is used to evaluate the complexity of the pre-processed surgical area image to be analyzed and obtain the image complexity score;

[0061] In the encoder layer, M residual blocks are included, each of which is composed of N convolutional layers and 1 residual connection; it is used to extract features of the pre-processed image of the art area to be analyzed, and obtain a feature image of the pre-processed art area to be analyzed;

[0062] Using the formula The specific value of M is calculated, F represents the image complexity score, and M max Indicates the maximum value of the preset M, M min Indicates the minimum value of the preset M, Represents a floor operation; where M max Can be 5, M min It can be 2, and the specific value is set by professional technicians according to actual conditions;

[0063] Using the formula The specific value of N is calculated, F represents the image complexity score, N max Indicates the maximum value of preset N, N min Indicates the minimum value of the preset N; where N max Can be 3, N min It can be 1, and the specific value is set by professional technicians according to actual conditions;

[0064] In the skip connection layer, it is used to assign weights to the pre-processed feature image of the technical area to be analyzed, so as to obtain a feature weight image of the pre-processed technical area to be analyzed;

[0065] In the decoder layer, it is used to perform feature fusion on the pre-processed feature weight image of the surgical area to be analyzed to obtain the feature image of the surgical area to be identified;

[0066] In the label classification layer, it is used to perform feature recognition on the feature image of the surgical area to be identified and obtain the surgical area status signal;

[0067] The number of residual blocks and convolutional layers in the encoder are dynamically adjusted according to the image complexity evaluation to adapt to surgical area images of different complexities. High-complexity images use more feature extraction layers to enhance the ability to capture complex features; low-complexity images reduce computational overhead and improve processing efficiency; in the encoder layer, the design of the residual block alleviates the gradient vanishing problem and enhances the model's ability to learn deep features; multi-layer convolutional layers extract multi-scale features to ensure that complex structures and details are captured; the jump connection layer assigns weights to the feature map through the attention gate to highlight the features of the key area while reducing the interference of irrelevant information; in the decoder, through step-by-step upsampling and jump connections, low-level and high-level features are fused, detail information is retained, and the integration of global context is achieved, effectively improving the resolution and expression ability of the feature image of the surgical area to be identified; the attention gate of the jump connection layer effectively suppresses noise and irrelevant feature interference, improving the model's adaptability to complex surgical area scenes;

[0068] The specific operations of training the image feature recognition layer include:

[0069] Collecting several groups of surgical area image state recognition training samples; each group of surgical area image state recognition training samples contains a verified state signal and a surgical area image; combining several groups of surgical area image state recognition training samples to obtain a surgical area image state recognition training set;

[0070] Inputting the surgical area image state recognition training set into the image feature recognition layer for model training to obtain an initial image feature recognition layer; performing model evaluation on the initial image feature recognition layer, if the initial image feature recognition layer passes the model evaluation, the initial image feature recognition layer is used as the image feature recognition layer in the surgical area image analysis model of the energy surgical device; otherwise, continuing model training using the surgical area image state recognition training set;

[0071] By collecting real surgical area image state recognition training samples and iteratively optimizing the training process, this method can significantly improve the accuracy, robustness, and generalization ability of the image feature recognition layer, provide a reliable core component for the surgical area image analysis model of energy surgical equipment, and further enhance the ability of real-time monitoring and intelligent control during surgery.

[0072] Based on the analysis of the surgical area state signal and the energy surgical device waveform superposition optimization model, the energy surgical device use waveform amplitude signal is obtained; the use waveform parameters of the energy surgical device are changed according to the energy surgical device use waveform amplitude signal;

[0073] The waveform superposition optimization model of energy surgical equipment includes a signal feature analysis layer, a smooth waveform optimization layer and a waveform superposition output layer; wherein the smooth waveform optimization layer is used to smoothly optimize the frequency waveform set to be superimposed output from the signal feature analysis layer, and smoothly connect it with the waveform amplitude signal used by the original energy surgical equipment to optimize the comfort of the waveform change of the energy surgical equipment; the waveform amplitude signal used by the original energy surgical equipment represents the corresponding waveform parameters used by the current energy surgical equipment;

[0074] According to the actual state of the surgical area, the waveform parameters of the energy surgical equipment are dynamically adjusted to adapt to the surgical needs more accurately. By superimposing different waveform amplitude signals, the efficiency of cutting and coagulation is improved to meet the diverse needs of complex surgical areas. The smooth waveform optimization layer smoothes the superimposed frequency waveform and smoothly connects it with the original waveform amplitude signal to avoid the risk of local heat accumulation or tissue carbonization caused by waveform mutations, optimize the waveform change curve, and make the energy surgical equipment more stable during the cutting or coagulation process, reducing thermal damage to surrounding healthy tissues.

[0075] The signal feature analysis layer is used to perform waveform matching on the state signal of the operation area to obtain the frequency waveform set X to be superimposed; where X={X i |i=1,2,…,I},X i represents the frequency waveform to be superimposed in the frequency waveform set X to be superimposed;

[0076] The signal feature analysis layer is constructed based on the SVM model;

[0077] The smooth waveform optimization layer is used to output the signal of the superimposed frequency waveform set X to obtain the waveform amplitude signal used by the energy surgical device;

[0078] The waveform superposition output layer is used to output the waveform amplitude signal used by the energy surgical device;

[0079] The specific steps of training the signal feature analysis layer include:

[0080] Collecting several groups of signal waveform matching training samples; each group of signal waveform matching training samples contains state signal features and corresponding frequency waveforms; combining several groups of signal waveform matching training samples to obtain a signal waveform matching training set;

[0081] The signal waveform matching training set is input into the SVM model for model training to obtain an initial signal feature analysis layer; the initial signal feature analysis layer is model evaluated, and if the initial signal feature analysis layer passes the model evaluation, the initial signal feature analysis layer is used as the signal feature analysis layer in the energy surgical device waveform superposition optimization model; otherwise, the signal waveform matching training set is used to continue model training;

[0082] The state signal characteristics and corresponding frequency waveforms are classified and mapped through the SVM model to accurately match the frequency waveform required by the surgical area, thereby improving the adaptability of the energy surgical equipment. The set of frequency waveforms to be superimposed output by the signal feature analysis layer can cover the complex state requirements of the surgical area and achieve accurate frequency waveform matching for different surgical scenarios. The smooth waveform optimization layer smoothes the frequency waveforms to be superimposed to ensure the continuity of the frequency waveform adjustment process and avoid thermal damage to local tissues caused by waveform mutations. The waveform amplitude signal used by the energy surgical equipment after smooth optimization is more stable, thereby improving the safety of intraoperative cutting and coagulation.

[0083] The specific steps of signal output in the smooth waveform optimization layer include:

[0084] Obtain the waveform amplitude signal used by the original energy surgical device; construct K smoothed wave signal individuals G based on the waveform amplitude signal used by the original energy surgical device and the frequency waveform set X to be superimposed k , k=1,2,…,K; smooth variable wave signal individual G k represents a variable wave superposition signal for energy surgical equipment; K smooth variable wave signal individuals G k Combine to obtain the individual iterative population of the smooth variable wave signal; set the maximum number of iterations; the maximum number of iterations is set by professional technicians according to actual conditions;

[0085] Construct a simulated signal wave variation evaluation model to extract the characteristics of the smooth wave variation signal and obtain the characteristics of the smooth wave variation signal; evaluate the comfort level of the smooth wave variation signal and obtain the comfort level of the simulated smooth wave variation signal; use the comfort level of the simulated smooth wave variation signal as the individual smooth wave variation signal G k The fitness S k ;

[0086] The population iteration process is divided into the first population iteration stage and the second population iteration stage;

[0087] In the first stage of population iteration, the smooth wave signal individual with the largest current fitness is selected as the elite smooth wave signal individual in each iteration. When the population is updated, the population is iteratively updated according to the elite smooth wave signal individuals to obtain a new smooth wave signal individual iterative population;

[0088] In the second stage of population iteration, a population mutation factor is introduced; in each iteration, the smooth wave signal individual iteration population is mutated and updated based on the population mutation factor to obtain a new smooth wave signal individual iteration population; the population mutation factor is set based on the Cauchy mutation strategy;

[0089] When the maximum number of iterations is reached, the smooth wave signal individual corresponding to the maximum fitness is output, which is the optimal smooth wave signal individual; based on the optimal smooth wave signal individual, the waveform amplitude signal used by the energy surgical device is output;

[0090] By optimizing the combination of waveform amplitude signals and frequency waveform sets to be superimposed on the original energy surgical equipment, the optimal smooth wave-changing signal that meets the needs of the surgical area is constructed. The signal wave-changing evaluation model and comfort evaluation mechanism are simulated to ensure that the generated signal can accurately adapt to the needs of different surgical area states. By constructing smooth wave-changing signal individuals, the continuity of the waveform amplitude signal during switching is ensured to avoid thermal damage or instability in the surgical area caused by waveform mutations. Comfort evaluation is introduced in the population iteration and optimization process, and the generated signal has higher smoothness, which helps to improve the safety of the surgical process.

[0091] The elite individual screening mechanism selects the signal individual with the largest fitness in each iteration, quickly converges to the optimal solution, improves the optimization efficiency, introduces the Cauchy mutation strategy for population mutation, breaks the local optimal limit, further explores the possibility of the optimal solution, and improves the globality of the optimization results; in the second stage of population mutation, the mutation strategy based on the Cauchy distribution enhances the breadth and diversity of the search space and reduces the risk of falling into the local optimal;

[0092] The optimal smooth wave variable signal ensures that the energy surgical equipment is more efficient during the cutting and coagulation process, reduces the operation time, automatically generates the optimal smooth wave variable signal, and reduces the burden of doctors manually adjusting the frequency waveform. The optimal smooth wave variable signal reduces the risk of tissue overheating or over-cutting and reduces postoperative complications. The smoothed and optimized signal is more suitable for the processing of complex tissues in the surgical area and protects the surrounding healthy tissues.

[0093] Example 2, a waveform superposition optimization system for energy surgical equipment, see Figure 1 As shown, including:

[0094] The energy surgery equipment surgical area recognition module includes an image acquisition unit and a feature analysis unit; the image acquisition unit is used to receive the surgical area image to be analyzed corresponding to the energy surgery equipment; the feature analysis unit is used to input the surgical area image to be analyzed into the energy surgery equipment surgical area image analysis model for analysis to obtain the surgical area state signal; the energy surgery equipment surgical area image analysis model includes an edge cutting layer, an image feature recognition layer and a surgical area state output layer, wherein the image feature recognition layer is used to perform feature recognition on the surgical area image to be analyzed based on the adaptively changed convolution layer and the self-attention mechanism;

[0095] The energy surgical equipment waveform superposition module includes a waveform analysis unit; the waveform analysis unit is used to analyze based on the surgical area state signal and the energy surgical equipment waveform superposition optimization model to obtain the energy surgical equipment use waveform amplitude signal; according to the energy surgical equipment use waveform amplitude signal, the use waveform parameters of the energy surgical equipment are changed; in the energy surgical equipment waveform superposition optimization model, a signal feature analysis layer, a smooth waveform optimization layer and a waveform superposition output layer are included;

[0096] The energy surgical equipment smooth wave changing module includes a superposition optimization unit; the superposition optimization unit is used to construct a smooth waveform optimization layer, the smooth waveform optimization layer is used to smoothly optimize the frequency waveform set to be superimposed output in the signal feature analysis layer, and smoothly connect it with the waveform amplitude signal used by the original energy surgical equipment to optimize the comfort of the waveform change of the energy surgical equipment; the waveform amplitude signal used by the original energy surgical equipment represents the corresponding waveform parameters used by the current energy surgical equipment.

[0097] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all such improvements and changes should fall within the scope of protection of the appended claims of the present invention. Parts not described in detail in this specification belong to the prior art known to those skilled in the art.

Claims

1. A waveform superposition optimization method for energy surgical equipment, characterized in that: The following steps are involved: Receiving an image of the surgical area to be analyzed corresponding to the energy surgical device; inputting the image of the surgical area to be analyzed into an image analysis model of the surgical area of ​​the energy surgical device for analysis to obtain a surgical area status signal; The energy surgery equipment surgical area image analysis model includes an edge cutting layer, an image feature recognition layer and a surgical area state output layer, wherein the image feature recognition layer is used to perform feature recognition on the surgical area image to be analyzed based on the adaptively changed convolution layer and the self-attention mechanism; Based on the analysis of the surgical area status signal and the energy surgical equipment waveform superposition optimization model, the energy surgical equipment waveform amplitude signal is obtained; the energy surgical equipment waveform parameters are changed according to the energy surgical equipment waveform amplitude signal; the energy surgical equipment waveform superposition optimization model includes a signal feature analysis layer, a smooth waveform optimization layer and a waveform superposition output layer; wherein the smooth waveform optimization layer is used to smoothly optimize the frequency waveform set to be superimposed output from the signal feature analysis layer, and smoothly connect it with the original energy surgical equipment waveform amplitude signal to optimize the comfort of the energy surgical equipment waveform change; the original energy surgical equipment waveform amplitude signal represents the corresponding waveform parameters of the current energy surgical equipment; The waveform superposition optimization model of energy surgical equipment includes a signal feature analysis layer, a smooth waveform optimization layer and a waveform superposition output layer; The signal feature analysis layer is used to perform waveform matching on the state signal of the operation area to obtain the frequency waveform set X to be superimposed; where X={X i |i=1,2,…,I},X i represents the frequency waveform to be superimposed in the frequency waveform set X to be superimposed; The signal feature analysis layer is constructed based on the SVM model; The smooth waveform optimization layer is used to output the signal of the superimposed frequency waveform set X to obtain the waveform amplitude signal used by the energy surgical device; The waveform superposition output layer is used to output the waveform amplitude signal used by the energy surgical device; The specific steps of signal output in the smooth waveform optimization layer include: Obtain the waveform amplitude signal used by the original energy surgical device; construct K smoothed wave signal individuals G based on the waveform amplitude signal used by the original energy surgical device and the frequency waveform set X to be superimposed k , k=1,2,…,K; smooth variable wave signal individual G k represents a variable wave superposition signal for energy surgical equipment; K smooth variable wave signal individuals G k Combine to obtain the individual iterative population of the smooth variable wave signal; set the maximum number of iterations; Construct a simulated signal wave variation evaluation model to extract the characteristics of the smooth wave variation signal and obtain the characteristics of the smooth wave variation signal; evaluate the comfort level of the smooth wave variation signal and obtain the comfort level of the simulated smooth wave variation signal; use the comfort level of the simulated smooth wave variation signal as the individual smooth wave variation signal G k The fitness S k ; The population iteration process is divided into the first population iteration stage and the second population iteration stage; In the first stage of population iteration, the smooth wave signal individual with the largest current fitness is selected as the elite smooth wave signal individual in each iteration. When the population is updated, the population is iteratively updated according to the elite smooth wave signal individuals to obtain a new smooth wave signal individual iterative population; In the second stage of population iteration, a population mutation factor is introduced; in each iteration, the smooth wave signal individual iteration population is mutated and updated based on the population mutation factor to obtain a new smooth wave signal individual iteration population; the population mutation factor is set based on the Cauchy mutation strategy; When the maximum number of iterations is reached, the smooth wave signal individual corresponding to the maximum fitness is output, which is the optimal smooth wave signal individual; based on the optimal smooth wave signal individual, the waveform amplitude signal used by the energy surgical device is output.

2. The method for optimizing waveform superposition of energy surgical equipment according to claim 1, characterized in that: The image analysis model of the surgical area of ​​energy surgical equipment includes an edge cutting layer, an image feature recognition layer and an surgical area status output layer; The edge cutting layer is used to perform edge cutting on the image of the surgical area to be analyzed, so as to obtain a pre-processed image of the surgical area to be analyzed; The image feature recognition layer is used to perform feature recognition on the pre-processed surgical area image to be analyzed to obtain the surgical area status signal; The image feature recognition layer is built based on the U-net architecture and transformer model; The surgical area status output layer is used to output the surgical area status signal.

3. The method for optimizing waveform superposition of energy surgical equipment according to claim 2, characterized in that: The image feature recognition layer includes a complexity calculation layer, an encoder layer, a decoder layer, a skip connection layer, and a label classification layer; Among them, the jump connection layer contains an attention gate; In the complexity calculation layer, it is used to evaluate the complexity of the pre-processed surgical area image to be analyzed and obtain the image complexity score; In the encoder layer, M residual blocks are included, each of which is composed of N convolutional layers and 1 residual connection; it is used to extract features of the pre-processed image of the art area to be analyzed, and obtain a feature image of the pre-processed art area to be analyzed; Using the formula The specific value of M is calculated, F represents the image complexity score, and M max Indicates the maximum value of the preset M, M min Indicates the minimum value of the preset M, Indicates floor operation; Using the formula The specific value of N is calculated, F represents the image complexity score, N max Indicates the maximum value of preset N, N min Indicates the minimum value of preset N; In the skip connection layer, it is used to assign weights to the pre-processed feature image of the area to be analyzed, so as to obtain a feature weight image of the pre-processed area to be analyzed; In the decoder layer, it is used to perform feature fusion on the pre-processed feature weight image of the surgical area to be analyzed to obtain the feature image of the surgical area to be identified; In the label classification layer, it is used to perform feature recognition on the feature image of the surgical area to be identified and obtain the surgical area status signal.

4. The method for optimizing waveform superposition of energy surgical equipment according to claim 3, characterized in that: The specific operations of training the image feature recognition layer include: Collecting several groups of surgical area image state recognition training samples; each group of surgical area image state recognition training samples contains a verified state signal and a surgical area image; combining several groups of surgical area image state recognition training samples to obtain a surgical area image state recognition training set; The surgical area image state recognition training set is input into the image feature recognition layer for model training to obtain the initial image feature recognition layer; the initial image feature recognition layer is subjected to model evaluation. If the initial image feature recognition layer passes the model evaluation, the initial image feature recognition layer is used as the image feature recognition layer in the surgical area image analysis model of the energy surgical device; otherwise, the surgical area image state recognition training set is used to continue model training.

5. The method for optimizing waveform superposition of energy surgical equipment according to claim 4, characterized in that: The specific steps of training the signal feature analysis layer include: Collecting several groups of signal waveform matching training samples; each group of signal waveform matching training samples contains state signal features and corresponding frequency waveforms; combining several groups of signal waveform matching training samples to obtain a signal waveform matching training set; The signal waveform matching training set is input into the SVM model for model training to obtain the initial signal feature analysis layer; the initial signal feature analysis layer is evaluated, and if the initial signal feature analysis layer passes the model evaluation, the initial signal feature analysis layer is used as the signal feature analysis layer in the waveform superposition optimization model of the energy surgical device; otherwise, the signal waveform matching training set is used to continue model training.

6. A waveform superposition optimization system for energy surgical equipment, characterized in that: The system applies a method for optimizing waveform superposition of an energy surgical device as described in any one of claims 1 to 5, including: The energy surgery equipment surgical area recognition module includes an image acquisition unit and a feature analysis unit; the image acquisition unit is used to receive the surgical area image to be analyzed corresponding to the energy surgery equipment; the feature analysis unit is used to input the surgical area image to be analyzed into the energy surgery equipment surgical area image analysis model for analysis to obtain the surgical area state signal; the energy surgery equipment surgical area image analysis model includes an edge cutting layer, an image feature recognition layer and a surgical area state output layer, wherein the image feature recognition layer is used to perform feature recognition on the surgical area image to be analyzed based on the adaptively changed convolution layer and the self-attention mechanism; The energy surgical equipment waveform superposition module includes a waveform analysis unit; the waveform analysis unit is used to analyze based on the surgical area state signal and the energy surgical equipment waveform superposition optimization model to obtain the energy surgical equipment use waveform amplitude signal; according to the energy surgical equipment use waveform amplitude signal, the use waveform parameters of the energy surgical equipment are changed; in the energy surgical equipment waveform superposition optimization model, a signal feature analysis layer, a smooth waveform optimization layer and a waveform superposition output layer are included; The energy surgical equipment smooth wave changing module includes a superposition optimization unit; the superposition optimization unit is used to construct a smooth waveform optimization layer, the smooth waveform optimization layer is used to smoothly optimize the frequency waveform set to be superimposed output in the signal feature analysis layer, and smoothly connect it with the waveform amplitude signal used by the original energy surgical equipment to optimize the comfort of the waveform change of the energy surgical equipment; the waveform amplitude signal used by the original energy surgical equipment represents the corresponding waveform parameters used by the current energy surgical equipment.

Images