Active intraoperative visual nerve monitoring method and system

Abstract

The invention discloses an active intraoperative visual nerve monitoring method and system, and relates to the technical field of nerve monitoring, and the method comprises the steps: obtaining a real-time motion image of a target organ receiving electrical stimulation; splitting the real-time motion image according to a preset timestamp to obtain a plurality of neural motion image segments; feature extraction is carried out on the neural motion image segments to obtain neural motion feature quantities, and the neural motion feature quantities comprise time sequence feature quantities and spatial feature quantities; inputting the neural motion characteristic quantity into a neural monitoring model for neural motion characteristic monitoring to obtain neural monitoring data; wherein the nerve monitoring model is obtained by training a preset neural network model according to a historical organ neural motion characteristic quantity set and a target labeling parameter. According to the scheme provided by the invention, an active and visual scheme is adopted, so that more accurate active monitoring capability with stronger purposiveness is provided, and the accuracy of intraoperative nerve detection is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of nerve monitoring, and in particular to an active intraoperative visual nerve monitoring method and system. Background Art

[0002] During medical surgery, the patient's nervous system is usually monitored to see if it is functioning normally through electromyographic signal acquisition, video image acquisition, and manual observation, in order to rule out the possibility of damage to the nervous system and organ tissue functions involved in the surgery. There are two main types of existing technical solutions. The first type is to collect electromyographic signals by attaching electrodes to the surface of the organ, and amplify and process them to form an electromyogram waveform displayed on an oscilloscope, so that professional medical personnel can analyze the electromyographic signals based on the waveform and make a judgment on whether the nervous system is working normally. The second type is to observe and monitor by video through equipment such as video laryngeal masks and endoscopes, using manual observation methods, with medical personnel observing the movement of organs through video monitoring terminals. Existing technical solutions are all passive methods, and the judgment conclusions rely on the experience and ability of medical personnel, which are prone to misjudgment and inaccurate judgment results. Summary of the Invention

[0003] In view of the above, the present invention aims to provide an active intraoperative visual nerve monitoring method and system to solve the above-mentioned technical problems.

[0004] The technical solution adopted in the present invention is as follows:

[0005] The present invention provides an active intraoperative visual nerve monitoring method, comprising:

[0006] Acquire real-time motion images of target organs receiving electrical stimulation;

[0007] Splitting the real-time motion image according to preset timestamps to obtain multiple neural motion image segments;

[0008] Extracting features from the neural motion image segment to obtain neural motion feature quantities, wherein the neural motion feature quantities include temporal feature quantities and spatial feature quantities;

[0009] The neuromotor feature quantity is input into a neuromotor feature monitoring model to obtain neuromotor feature data; wherein the neuromotor feature monitoring model is obtained by training a preset neural network model according to a historical organ neuromotor feature quantity set and target labeling parameters.

[0010] Optionally, the real-time motion image is split according to preset time stamps to obtain multiple neural motion image segments, including:

[0011] Based on a preset time stamp, the real-time motion image is time-aligned with the electrical stimulation frequency to obtain a neural motion image segment.

[0012] Optionally, feature extraction is performed on the neural motion image segment to obtain neural motion feature quantities, including:

[0013] Segmenting the neural motion image segment into a plurality of time-series windows, each time-series window containing a preset number of frames of images;

[0014] Sliding sampling is performed in each of the time series windows according to a preset step size to obtain a space-time cube;

[0015] According to the space-time cube, a neural motion feature quantity is obtained.

[0016] Optionally, obtaining a neural motion feature according to the space-time cube includes:

[0017] Performing one-dimensional convolution processing on the space-time cube to obtain a time series feature quantity;

[0018] Performing two-dimensional convolution processing on the space-time cube to obtain spatial feature quantities;

[0019] The temporal feature quantity and the spatial feature quantity are cross-stacked to obtain a neuromotor feature quantity.

[0020] Optionally, the neuromotor feature is input into a neuromotor feature monitoring model to perform neuromotor feature monitoring to obtain neuromotor monitoring data, including:

[0021] Determining the type of surgery, the name of the target organ, and the position change of the target organ's motion feature points based on the neuromotor feature quantity;

[0022] The nerve monitoring data is obtained according to the type of surgery, the name of the target organ, and the position change of the target organ's motion characteristic point.

[0023] Optionally, the training process of the neural monitoring model includes:

[0024] Acquire a historical organ neural motion feature value set, wherein the historical organ neural motion feature value set includes a historical organ neural motion time series feature value set and a historical organ neural motion space feature value set;

[0025] Labeling the historical organ neuromotor feature quantity set to obtain the historical organ neuromotor feature quantity training set;

[0026] The preset neural network model is trained using the historical organ neural motion feature training set and the cross entropy loss function to obtain a neural monitoring model.

[0027] Optionally, the historical organ neuromotor feature set is labeled to obtain a historical organ neuromotor feature training set, including:

[0028] Each feature quantity in the historical organ neural motion feature quantity set is labeled with surgery type parameters, motion feature point parameters, motion feature line parameters, and motion feature trajectory parameters to obtain a historical organ neural motion feature quantity training set.

[0029] Optionally, labeling the historical organ neuromotor feature set to obtain a historical organ neuromotor feature training set also includes:

[0030] The annotated historical organ neuromotor feature set is divided into 80%, 15%, and 5%, where 80% of the data is used as the training data set, 15% of the data is used as the test set, and 5% of the data is used as the sample set to train the neural monitoring model and verify the pre-monitoring effect.

[0031] The present invention also provides an active intraoperative visual nerve monitoring system for executing the above method, the system comprising:

[0032] An electric pulse nerve stimulation device, an image acquisition device, a server, and a video monitoring device, wherein the electric pulse nerve stimulation device and the image acquisition device are electrically connected to the server, and the server is electrically connected to the video monitoring device;

[0033] The electric pulse nerve stimulation device is used to electrically stimulate the target organ;

[0034] The image acquisition device is used to obtain real-time motion images of the target organ receiving electrical stimulation;

[0035] The server is used to receive the electrical stimulation frequency of the electrical pulse nerve stimulation device and the real-time motion image of the image acquisition device, and output nerve monitoring data;

[0036] The video monitoring device is used to receive the nerve monitoring data output by the server and display the real-time motion image and nerve monitoring data of the target organ receiving electrical stimulation.

[0037] Optionally, a speaker is provided inside the video surveillance device, and when the nerve monitoring data is abnormal, the speaker sends an alarm signal.

[0038] The above solution of the present invention includes at least the following beneficial effects:

[0039] The above-mentioned solution of the present invention obtains real-time motion images of the target organ receiving electrical stimulation; splits the real-time motion images according to preset timestamps to obtain multiple neural motion image segments; extracts features from the neural motion image segments to obtain neural motion feature quantities, which include temporal feature quantities and spatial feature quantities; inputs the neural motion feature quantities into a neural monitoring model to perform neural motion feature monitoring to obtain neural monitoring data; wherein the neural monitoring model is obtained by training a preset neural network model based on a historical organ neural motion feature quantity set and target annotation parameters. The solution of the present invention provides a more accurate and more targeted active monitoring capability with an active and visual solution, thereby improving the accuracy of intraoperative neural detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be further described below with reference to the accompanying drawings, in which:

[0041] Figure 1 This is a flow chart of the active intraoperative visual nerve monitoring method provided by an embodiment of the present invention.

[0042] Figure 2 This is a schematic diagram of the structure of an active intraoperative visual nerve monitoring system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0043] The following describes embodiments of the present invention in detail. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended only to explain the present invention and are not to be construed as limiting the present invention.

[0044] The present invention proposes an embodiment of an active intraoperative visual nerve monitoring method, specifically, Figure 1 shown, including:

[0045] Step 11, obtaining real-time motion images of the target organ receiving electrical stimulation;

[0046] Step 12, splitting the real-time motion image according to preset timestamps to obtain multiple neural motion image segments;

[0047] Step 13, extracting features from the neural motion image segment to obtain neural motion feature quantities, wherein the neural motion feature quantities include temporal feature quantities and spatial feature quantities;

[0048] Step 14: Input the neuromotor feature into a neuromotor feature monitoring model to obtain neuromotor feature data; wherein the neuromotor feature is obtained by training a preset neural network model based on a historical organ neuromotor feature set and target labeling parameters.

[0049] In this embodiment, an image acquisition device must be deployed before surgery. Depending on the surgical scenario, a video laryngeal mask, endoscope, or external camera is selected and placed or deployed, confirming that it is communicating effectively with the server and video surveillance equipment. Then, during surgery, the electrical nerve stimulation device is activated, applying electrical pulses to the exposed nerves at the set current intensity and discharge frequency. The image acquisition device captures real-time images of the target organ or tissue and transmits the video data to the server. The server then uses a neural monitoring model tailored to the surgery, selects accurate video clips based on information such as timestamps, and reviews them frame by frame. The server then records the positional changes of the target organ's motion feature points and calculates a quantitative, comprehensive analysis result. The neural detection model here includes neural monitoring models for various surgical types and target organs, as well as the server's built-in algorithm. The back-end service running on the server receives real-time images transmitted by the image acquisition device. Its deep learning model first captures the temporal and spatial characteristics of the dynamic changes in the cavity or organ through a sliding cube sampling engine, and applies a decomposed 3D convolutional architecture to enhance feature expression. The trained recognition model infers the surgical type and target organ, loads the corresponding visual analysis algorithm model and calls it, making it more targeted and improving the accuracy of intraoperative neural monitoring. Then, based on the different surgical types and target organs, the comprehensive analysis results can be expressed as percentage of movement amplitude, percentage of contraction amplitude, vibration frequency, or other types of quantitative conclusions. Based on the result data, quantitative display information is drawn on the screen of the video surveillance device. If necessary, the built-in speaker can be activated according to preset rules to issue an audible alarm.

[0050] This embodiment adopts a visual solution, which makes the arrangement of the acquisition device more intuitive, improves effectiveness and stability, and reduces the error rate; it flexibly adopts multiple types of visual models, has sufficient flexibility and adaptability, and improves accuracy, reduces human intervention, and reduces the misjudgment rate; with an active and visual solution, it provides more precise and more purposeful active monitoring capabilities, thereby improving the accuracy of intraoperative nerve detection.

[0051] In an optional embodiment of the present invention, step 12 may include:

[0052] Based on a preset time stamp, the real-time motion image is time-aligned with the electrical stimulation frequency to obtain a neural motion image segment.

[0053] In this embodiment, medical personnel operate the electric probe of the electric pulse nerve stimulation device to apply electric pulse stimulation to the exposed nerve. When the discharge is completed, the electric pulse nerve stimulation device synchronously sends data such as the stimulation current intensity, frequency, frequency, and timestamp to the server;

[0054] After receiving video data from the image acquisition device, the server splits the video data into multiple frames according to the time sequence and then aligns the stimulation current frequency with the image frame time. This temporal alignment of the electrical stimulation frequency and the video data provides a precise monitoring cycle, reduces interference rates, and improves analysis accuracy.

[0055] In an optional embodiment of the present invention, step 13 may include:

[0056] Step 131 , dividing the neural motion image segment into a plurality of time-series windows, each time-series window containing a preset number of frames of images;

[0057] Step 132: Sliding sampling is performed in each of the time series windows according to a preset step size to obtain a space-time cube;

[0058] According to the space-time cube, a neural motion feature quantity is obtained.

[0059] In this embodiment, the video image data of the image acquisition device is normalized and divided into multiple time windows containing 16 frames of images. Within each time window, sliding window sampling is performed with a spatial step of 32×32 pixels and a temporal step of 4 frames, and a space-time cube is extracted.

[0060] A 4×4 two-dimensional convolution kernel is used to extract spatial features of the image frame by frame to obtain spatial feature quantities; a 4×1×1 one-dimensional convolution kernel is used to extract dynamic features along the time dimension to obtain temporal feature quantities, which include not only time features but also the current (electrical stimulation) intensity features at the current moment; spatial convolution and temporal convolution are cross-stacked to construct a multi-layer decomposition 3D convolution network, which ultimately completes the highlighting of key features and the recognition of organ motion features, and is mapped into neural motion feature quantities through a fully connected layer.

[0061] In an optional embodiment of the present invention, step 14, inputting the neuromotor feature into a neuromotor feature monitoring model to obtain neuromotor feature monitoring data may include:

[0062] Step 141, determining the type of surgery, the name of the target organ, and the position change of the target organ motion feature point based on the neuromotor feature quantity;

[0063] Step 142 : Obtaining nerve monitoring data based on the type of surgery, the name of the target organ, and the position change of the target organ's motion feature point.

[0064] In this embodiment, after the neural motion features are input into the neural monitoring model, the neural monitoring model analyzes the neural motion features to obtain parameters such as the surgery type, target organ name, and the position change of the target organ's motion feature points. The time series decoder then restores the feature point position changes and reconstructs the motion point trajectory. The various parameter data are then fused to calculate the mechanical index of the neural motion:

[0065] Neuromotor range:

[0066] ;

[0067] Assessment of nerve excitability: A <50%, weak response; A 50-100%, normal; A >100%, overexcited.

[0068] Where A is the amplitude of nerve movement; is the maximum offset distance of the feature point within the time window; is the initial coordinate of the feature point at the beginning of electrical stimulation; is the coordinate of the feature point at time t; t is time; is the Euclidean distance, which quantifies the spatial displacement; is the standard size of the target organ in the image.

[0069] Nerve contraction frequency:

[0070] ;

[0071] ;

[0072] ;

[0073] Assess nerve conduction function: <2Hz, conduction block; =2-5Hz, normal; >5Hz, spasm.

[0074] in, is the nerve contraction frequency; is the displacement change; is the time window; N is the number of feature points; For zero-crossing detection, capture the periodic contraction of the nerve; is the sampling interval (determine the interval time according to the timestamp), is the indicator function, is a symbolic function.

[0075] In an optional embodiment of the present invention, the training process of the neural monitoring model includes:

[0076] Acquire a historical organ neural motion feature value set, wherein the historical organ neural motion feature value set includes a historical organ neural motion time series feature value set and a historical organ neural motion space feature value set;

[0077] Labeling the historical organ neuromotor feature quantity set to obtain the historical organ neuromotor feature quantity training set;

[0078] The preset neural network model is trained using the historical organ neural motion feature training set and the cross entropy loss function to obtain a neural monitoring model.

[0079] In this embodiment, professional medical personnel first collect a large number of organ motion images taken by endoscopes or external cameras, and complete the annotation of content such as surgery type, motion feature points, motion feature lines, and motion feature trajectories, which serve as training sets, test sets, and sample sets for model training.

[0080] Using the labeled imagery, supervised training of the convolutional neural network model is performed using either cross-entropy or Dice loss. Data augmentation, regularization, and early stopping strategies are incorporated to optimize model parameters and prevent overfitting. Through continuous infusion of imagery, manual parameter optimization, and model self-optimization, the model's recognition and judgment performance ultimately meets the target, completing training.

[0081] The acquisition of historical organ neuromotor feature training sets includes:

[0082] Each feature quantity in the historical organ neural motion feature quantity set is labeled with surgery type parameters, motion feature point parameters, motion feature line parameters, and motion feature trajectory parameters to obtain a historical organ neural motion feature quantity training set.

[0083] The annotated historical organ neuromotor feature set is divided into 80%, 15%, and 5%, where 80% of the data is used as the training data set, 15% of the data is used as the test set, and 5% of the data is used as the sample set to train the neural monitoring model and verify the pre-monitoring effect.

[0084] An embodiment of the present invention further provides an active intraoperative visual nerve monitoring system for executing the above method, the system comprising:

[0085] An electric pulse nerve stimulation device, an image acquisition device, a server, and a video monitoring device, wherein the electric pulse nerve stimulation device and the image acquisition device are electrically connected to the server, and the server is electrically connected to the video monitoring device;

[0086] The electric pulse nerve stimulation device is used to electrically stimulate the target organ;

[0087] The image acquisition device is used to obtain real-time motion images of the target organ receiving electrical stimulation;

[0088] The server is used to receive the electrical stimulation frequency of the electrical pulse nerve stimulation device and the real-time motion image of the image acquisition device, and output nerve monitoring data;

[0089] The video monitoring device is used to receive the nerve monitoring data output by the server and display the real-time motion image and nerve monitoring data of the target organ receiving electrical stimulation.

[0090] In this embodiment, Figure 2 As shown in the figure, this system consists of main hardware such as image acquisition device, electric pulse nerve stimulation equipment, server, video surveillance equipment, and main software such as server built-in model algorithm and supporting app application.

[0091] Image acquisition devices: such as endoscopes, video laryngeal masks, external high-definition cameras and other equipment. Depending on the type of surgery, different devices are used to collect real-time images of the target organs or tissues and transmit the video data to the server.

[0092] Electric pulse nerve stimulation equipment: Parameters such as stimulation current, frequency, and frequency can be set. External devices such as electric probes and electrodes extending from the equipment actively apply electric pulse stimulation to targets such as nerves or skin tissues exposed during surgery, forming nerve impulses and conducting them to control organ muscle movement.

[0093] Server: Serves big data analysis work, loads visual analysis models for different surgical scenarios, different target organs and different electrical pulse patterns, performs qualitative and quantitative analysis on the video sent back by the video acquisition device based on the configuration parameters and timestamp data sent back by the electrical pulse nerve stimulation device, and transmits the analysis result data together with the video image to the video surveillance device for display.

[0094] The video surveillance device can be a PC monitor, portable Pad or other device. It displays the video images and analysis result data received from the server through a pre-installed app application, and draws quantitative display information on the video screen based on the result data. The video surveillance device is equipped with a speaker inside, and when necessary, the built-in speaker can be activated according to preset rules to issue a sound alarm.

[0095] Specifically, the server sends the comprehensive analysis results and video clip stream data to the video surveillance device. The front-end app plots the comprehensive analysis results in a quantitative display format and overlays them with the video clips. If the comprehensive analysis results contain warning information such as nerve damage, the speaker activates and sounds an audible alarm according to the alarm rules.

[0096] It should be noted that the system is a system corresponding to the above-mentioned method, and all implementation methods in the above-mentioned method embodiments are applicable to the embodiments of the system and can achieve the same technical effects.

[0097] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0098] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0099] In the embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be through some interface, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0100] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0101] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0102] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, ROM, RAM, a magnetic disk, or an optical disk.

[0103] In addition, it should be pointed out that in the apparatus and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent solutions of the present invention. Moreover, the steps of performing the above-mentioned series of processing can naturally be performed in chronological order according to the order of description, but they do not necessarily need to be performed in chronological order, and some steps can be performed in parallel or independently of each other. For those of ordinary skill in the art, it can be understood that all or any steps or components of the method and apparatus of the present invention can be implemented in hardware, firmware, software or a combination thereof in any computing device (including a processor, storage medium, etc.) or a network of computing devices. This can be achieved by those of ordinary skill in the art using their basic programming skills after reading the description of the present invention.

[0104] Therefore, the purpose of the present invention can also be achieved by running a program or a group of programs on any computing device. The computing device can be a well-known general-purpose device. Therefore, the purpose of the present invention can also be achieved simply by providing a program product containing program code that implements the method or device. That is to say, such a program product also constitutes the present invention, and the storage medium storing such a program product also constitutes the present invention. Obviously, the storage medium can be any well-known storage medium or any storage medium developed in the future. It should also be pointed out that in the device and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. In addition, the steps of performing the above-mentioned series of processing can naturally be performed in chronological order according to the order of description, but do not necessarily need to be performed in chronological order. Certain steps can be performed in parallel or independently of each other.

[0105] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. An active intraoperative visual nerve monitoring method, characterized in that: include: Acquire real-time motion images of target organs receiving electrical stimulation; Splitting the real-time motion image according to preset timestamps to obtain multiple neural motion image segments; Extracting features from the neural motion image segment to obtain neural motion feature quantities, wherein the neural motion feature quantities include temporal feature quantities and spatial feature quantities; The neuromotor feature quantity is input into a neuromotor feature monitoring model to obtain neuromotor feature data; wherein the neuromotor feature monitoring model is obtained by training a preset neural network model according to a historical organ neuromotor feature quantity set and target labeling parameters.

2. The active intraoperative visual nerve monitoring method according to claim 1, characterized in that: The real-time motion image is split according to preset time stamps to obtain multiple neural motion image segments, including: Based on a preset time stamp, the real-time motion image is time-aligned with the electrical stimulation frequency to obtain a neural motion image segment.

3. The active intraoperative visual nerve monitoring method according to claim 1, characterized in that: Extracting features from the neural motion image segment to obtain neural motion feature quantities includes: Segmenting the neural motion image segment into a plurality of time-series windows, each time-series window containing a preset number of frames of images; Sliding sampling is performed in each of the time series windows according to a preset step size to obtain a space-time cube; According to the space-time cube, a neural motion feature quantity is obtained.

4. The active intraoperative visual nerve monitoring method according to claim 3, characterized in that: According to the space-time cube, the neural motion feature quantity is obtained, including: Performing one-dimensional convolution processing on the space-time cube to obtain a time series feature quantity; Performing two-dimensional convolution processing on the space-time cube to obtain spatial feature quantities; The temporal feature quantity and the spatial feature quantity are cross-stacked to obtain a neuromotor feature quantity.

5. The active intraoperative visual nerve monitoring method according to claim 1, characterized in that: Inputting the neuromotor feature into a neuromotor monitoring model to monitor the neuromotor feature, and obtaining neuromotor monitoring data, including: Determining the type of surgery, the name of the target organ, and the position change of the target organ's motion feature points based on the neuromotor feature quantity; The nerve monitoring data is obtained according to the type of surgery, the name of the target organ, and the position change of the target organ's motion characteristic point.

6. The active intraoperative visual nerve monitoring method according to claim 1, characterized in that: The training process of the neural monitoring model includes: Acquire a historical organ neural motion feature value set, wherein the historical organ neural motion feature value set includes a historical organ neural motion time series feature value set and a historical organ neural motion space feature value set; Labeling the historical organ neuromotor feature quantity set to obtain the historical organ neuromotor feature quantity training set; The preset neural network model is trained using the historical organ neural motion feature training set and the cross entropy loss function to obtain a neural monitoring model.

7. The active intraoperative visual nerve monitoring method according to claim 6, characterized in that: The historical organ neuromotor feature set is labeled to obtain a historical organ neuromotor feature training set, including: Each feature quantity in the historical organ neural motion feature quantity set is labeled with surgery type parameters, motion feature point parameters, motion feature line parameters, and motion feature trajectory parameters to obtain a historical organ neural motion feature quantity training set.

8. The active intraoperative visual nerve monitoring method according to claim 6, characterized in that: The historical organ neuromotor feature set is labeled to obtain a historical organ neuromotor feature training set, which also includes: The annotated historical organ neuromotor feature set is divided into 80%, 15%, and 5%, where 80% of the data is used as the training data set, 15% of the data is used as the test set, and 5% of the data is used as the sample set to train the neural monitoring model and verify the pre-monitoring effect.

9. An active intraoperative visual nerve monitoring system, characterized in that: The method according to any one of claims 1 to 8 is performed, wherein the system comprises: An electric pulse nerve stimulation device, an image acquisition device, a server, and a video monitoring device, wherein the electric pulse nerve stimulation device and the image acquisition device are electrically connected to the server, and the server is electrically connected to the video monitoring device; The electric pulse nerve stimulation device is used to electrically stimulate the target organ; The image acquisition device is used to obtain real-time motion images of the target organ receiving electrical stimulation; The server is used to receive the electrical stimulation frequency of the electrical pulse nerve stimulation device and the real-time motion image of the image acquisition device, and output nerve monitoring data; The video monitoring device is used to receive the nerve monitoring data output by the server and display the real-time motion image and nerve monitoring data of the target organ receiving electrical stimulation.

10. The active intraoperative visual nerve monitoring system according to claim 9, characterized in that: The video monitoring device is internally provided with a speaker, and when the nerve monitoring data is abnormal, the speaker sends out an alarm signal.

Images