Intelligent gynecological operating bed control system

Abstract

The invention relates to the field of medical equipment control, in particular to an intelligent gynecological operating bed control system, and particularly relates to a system for realizing intelligent control of a gynecological operating bed by using a deep learning technology and an image processing technology. The system comprises a data acquisition module, a neural network processing module, a feature recognition module and a bed body control module. Firstly, a data acquisition module captures vaginal ultrasound and grayscale images; then, a neural network processing module receives and trains the images and constructs a feature classification model set; the feature recognition module performs classification prediction on a new vagina image by using the model set; finally, the bed body control module adjusts the position and posture of the operating bed according to the prediction result. According to the system, ultrasonic and grayscale images are synchronously processed through bimodal image analysis, the accuracy and reliability of feature recognition are remarkably improved, and an efficient and intelligent equipment control solution is provided for gynecological operations.

Description

Technical Field

[0001] The present invention relates to the field of medical device control, specifically to an intelligent control system for a gynecological operating table, and particularly to a system that realizes intelligent control of a gynecological operating table by using deep learning technology and image processing technology. Background Art

[0002] As an important part of clinical medicine, gynecological surgery has high requirements for the precise control and positioning of the operating table. Traditional gynecological operating tables usually adopt mechanical adjustment or simple electric control methods. Doctors need to manually adjust the position, angle and height of the table body, which not only increases the workload of doctors, but also may affect the surgical accuracy due to human operation errors.

[0003] The existing gynecological operating table control systems mainly have the following problems: First, they lack the ability to intelligently identify the patient's state and cannot automatically adjust the table body position according to the patient's physiological characteristics; second, the control accuracy is limited and it is difficult to meet the high requirements for precise positioning in minimally invasive surgery, etc.; third, they lack the remote control ability and are not convenient for realizing telemedicine and surgical guidance; finally, the coordination between systems is poor and they cannot effectively integrate image data and table body control.

[0004] With the development of artificial intelligence and computer vision technology, it has become possible to apply deep learning and image processing technology to the field of medical device control. However, at present, there is no system that can effectively integrate image recognition technology and table body control technology to realize intelligent control of a gynecological operating table.

[0005] Therefore, there is an urgent need for an intelligent control system for a gynecological operating table that can intelligently adjust according to the patient's physiological characteristics, has high-precision control ability, supports remote operation and has high coordination efficiency among various modules. Summary of the Invention

[0006] The purpose of the present invention is to provide an intelligent control system for a gynecological operating table, which realizes intelligent table body control based on the patient's physiological characteristics by integrating deep learning technology, image processing technology and precision mechanical control technology, and solves the problems such as insufficient control accuracy, lack of intelligence and remote control ability of the existing gynecological operating tables.

[0007] The present invention provides an intelligent control system for a gynecological operating table, including:

[0008] A data acquisition module, which is used to acquire vaginal ultrasound images and vaginal grayscale images;

[0009] A neural network processing module, which is communicatively connected to the data acquisition module, and is used to receive the vaginal ultrasound images and the vaginal grayscale images, and train a feature classification model set based on the vaginal ultrasound images and the vaginal grayscale images;

[0010] A feature recognition module, communicatively connected to the neural network processing module, for inputting a vaginal image to be predicted into the feature classification model set to obtain a feature classification result;

[0011] A bed control module, communicatively connected to the feature recognition module, for controlling the position and posture of a gynecological operating bed based on the feature classification result.

[0012] Preferably, the neural network processing module includes:

[0013] A first network construction unit, for training the vaginal ultrasound image using a first convolutional neural network to obtain a vaginal ultrasound image feature classification model;

[0014] A second network construction unit, for training the vaginal grayscale image using a second convolutional neural network to obtain a vaginal grayscale image feature classification model;

[0015] Wherein, the feature classification model set includes the vaginal ultrasound image feature classification model and the vaginal grayscale image feature classification model.

[0016] Preferably, the second network construction unit includes:

[0017] A first adjustment unit, for randomly adjusting the input order of the vaginal grayscale image data set;

[0018] A second adjustment unit, for adding interference data to the vaginal grayscale image data set during the training process;

[0019] A third adjustment unit, for updating the network node weight value of the second convolutional neural network after each training cycle ends.

[0020] Preferably, the feature recognition module includes:

[0021] A first calculation unit, for calculating a first confidence level of the vaginal ultrasound image feature classification result;

[0022] A second calculation unit, for calculating a second confidence level of the vaginal grayscale image feature classification result;

[0023] A feature fusion unit, for performing weighted fusion on the vaginal ultrasound image feature classification result and the vaginal grayscale image feature classification result based on the first confidence level and the second confidence level to obtain a final feature classification result.

[0024] Preferably, the bed control module includes:

[0025] A first adjustment unit, for receiving an adjustment instruction of the bed position and controlling the horizontal position and vertical position of the gynecological operating bed;

[0026] A second adjustment unit, configured to receive an adjustment instruction for the angle of the bed body and control the angle of the gynecological operating bed body;

[0027] A third adjustment unit, configured to receive an adjustment instruction for the height of the bed body and control the height of the gynecological operating bed body.

[0028] Preferably, the first adjustment unit includes:

[0029] A first adjustment motor, configured to provide driving force;

[0030] A rotating rod, connected to the first adjustment motor, configured to receive the driving force of the first adjustment motor and rotate;

[0031] A fixed rod, disposed at the bottom of the operating bed body and connected to the rotating rod, configured to drive the operating bed body to move under the drive of the rotating rod, so as to realize the adjustment of the horizontal position and vertical position of the operating bed body.

[0032] Preferably, it further includes:

[0033] A remote central control module, communicatively connected to the bed body control module, including:

[0034] An image processing unit, configured to process image information from an image acquisition device;

[0035] A command distribution unit, configured to generate a control command based on the image information and send it to the bed body control module;

[0036] An operation feedback unit, configured to receive the execution information of the bed body control module and perform evaluation processing.

[0037] Preferably, it further includes:

[0038] An auxiliary system, communicatively connected to the remote central control module, including:

[0039] An auxiliary bed, serving as an auxiliary tool for the operating bed;

[0040] A positioning module, disposed on the auxiliary bed, configured to obtain the position information of the auxiliary bed;

[0041] A position relationship determination unit, disposed in the remote central control module, configured to determine the position relationship between the auxiliary bed and the operating bed based on the position information.

[0042] Preferably, the remote central control module further includes:

[0043] A voice recognition unit, configured to recognize the content of voice information;

[0044] A voice spatial positioning unit, configured to determine the position of the voice information sender;

[0045] A voice command processing unit, configured to convert the voice information into a control command and send it to the bed control module to control the position or attitude of the operating bed.

[0046] Preferably, the remote central control module further includes:

[0047] A storage unit, configured to store surgical information, surgical video information, surgical operation sequence information, and video monitoring information;

[0048] An evaluation processing unit, configured to perform evaluation processing on the surgical operation based on the surgical information and the surgical video information.

[0049] The beneficial effects of the present invention include: First, through the dual-modal image analysis system, the vaginal ultrasound image and the vaginal grayscale image are processed simultaneously, significantly improving the accuracy and reliability of feature recognition; Second, an innovative adaptive neural network training method is adopted to enhance the robustness and generalization ability of the model; Third, based on the confidence-weighted feature fusion mechanism, the accuracy and reliability of the classification result are ensured; Finally, through the multi-degree-of-freedom bed adjustment system, precise control of the bed position and attitude is achieved, improving the surgical precision and safety. In addition, the present invention also supports remote control and multi-bed collaborative operation, providing technical support for telemedicine. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a schematic diagram of the overall structure of an intelligent gynecological operating bed control system of the present invention;

[0051] Figure 2 It is a schematic diagram of the structure of the neural network processing module of the present invention;

[0052] Figure 3 It is a schematic diagram of the structure of the feature recognition module of the present invention;

[0053] Figure 4 It is a schematic diagram of the structure of the bed control module of the present invention;

[0054] Figure 5 It is a schematic diagram of the structure of the remote central control module of the present invention;

[0055] Figure 6 It is a schematic diagram of the structure of the auxiliary system of the present invention;

[0056] Figure 7 It is a flowchart of the working process of the first adjustment unit of the present invention;

[0057] Figure 8 It is a flowchart of the working process of the second adjustment unit of the present invention;

[0058] Figure 9 The working flowchart of the third adjustment unit of the present invention;

[0059] Figure 10 The working flowchart of the feature fusion unit of the present invention. Specific embodiments

[0060] Please refer to the appendix Figures 1-10 and the present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0061] As Figure 1 shown, an intelligent gynecological operating bed control system provided by the present invention includes a data acquisition module 1, a neural network processing module 2, a feature recognition module 3, and a bed body control module 4.

[0062] The data acquisition module 1 is used to acquire vaginal ultrasound images and vaginal grayscale images. Preferably, the data acquisition module 1 includes an ultrasound image acquisition unit 11 and a grayscale image acquisition unit 12. Among them, the ultrasound image acquisition unit 11 can be a medical ultrasound probe for acquiring ultrasound images of the patient's vaginal area; the grayscale image acquisition unit 12 can be a high-definition camera for acquiring grayscale images of the patient's vaginal area.

[0063] In a specific embodiment of the present invention, the resolution of the ultrasound images acquired by the ultrasound image acquisition unit 11 is 1024×768 pixels, and the sampling frequency is 30 frames / second; the resolution of the grayscale images acquired by the grayscale image acquisition unit 12 is 2048×1536 pixels, and the sampling frequency is 60 frames / second. This configuration can ensure that the image quality meets the requirements of subsequent analysis and processing, while maintaining the real-time performance of the system.

[0064] The neural network processing module 2 is communicatively connected to the data acquisition module 1, and is used to receive vaginal ultrasound images and vaginal grayscale images, and train a feature classification model set based on these images. Specifically, as Figure 2 shown, the neural network processing module 2 includes a first network construction unit 21 and a second network construction unit 22.

[0065] The first network construction unit 21 is used to train vaginal ultrasound images using a first convolutional neural network to obtain a vaginal ultrasound image feature classification model. In a preferred embodiment of the present invention, the first convolutional neural network adopts a ResNet-50 architecture, including 50 convolutional layers, and the input image size is 224×224 pixels. The first network construction unit 21 first preprocesses the vaginal ultrasound images, including size adjustment, noise removal, and contrast enhancement, and then inputs the processed images into the first convolutional neural network for training. During the training process, the cross-entropy loss function is used, the optimizer adopts the Adam algorithm, the initial learning rate is set to 0.001, and the learning rate is reduced to 0.1 times the original every 30 training epochs.

[0066] The second network construction unit 22 is used to train vaginal grayscale images using a second convolutional neural network to obtain a vaginal grayscale image feature classification model. In another preferred embodiment of the present invention, the second convolutional neural network adopts a DenseNet-121 architecture, including 121 convolutional layers, and the input image size is 224×224 pixels. The second network construction unit 22 also preprocesses the vaginal grayscale images and then inputs the processed images into the second convolutional neural network for training.

[0067] The feature classification model set includes a vaginal ultrasound image feature classification model and a vaginal grayscale image feature classification model, and these two models together constitute the feature classification basis of the present invention.

[0068] As Figure 2 As further shown, the second network construction unit 22 includes a first adjustment unit 221, a second adjustment unit 222, and a third adjustment unit 223. These three adjustment units work together to optimize the training process and performance of the neural network.

[0069] The first adjustment unit 221 is used to randomly adjust the input order of the vaginal grayscale image dataset. Specifically, the first adjustment unit 221 calculates the training adjustment probability according to the following formula:

[0070]

[0071] Where, P adj is the input order adjustment probability of the vaginal grayscale image dataset; α is the training adjustment factor, usually set to a value between 0.5 and 1.0; R is the total number of input images in the vaginal grayscale image dataset; r is the number of input images in a training epoch; P in is the probability that the image to be input appears in the image set; P out is the probability that the already input image appears in the image set; N in (r) is the number of input images of the vaginal grayscale image dataset in the already used dataset before adjustment in the r-th training epoch; Nout (r) is the number of input images in the used dataset after adjustment of the vaginal grayscale image dataset in the r-th training cycle; δ is the adjustment interval of the image input data; N δ is the number of data corresponding to the adjustment interval in the r-th training cycle; N is the total number of images in the vaginal grayscale image dataset.

[0072] In a specific embodiment of the present invention, α is set to 0.75, R is 10000, r is 100, and δ is 10. These parameter values are the optimal configurations verified by a large number of experiments and can effectively balance the training efficiency and the model performance. The first adjustment unit 221 adjusts the input order of the vaginal grayscale image dataset according to the calculated adjustment probability, thereby optimizing the training process.

[0073] The second adjustment unit 222 is used to add interference data to the vaginal grayscale image dataset during the training process. Specifically, the second adjustment unit 222 includes an interference data addition subunit 2221 and a training processing subunit 2222.

[0074] The interference data addition subunit 2221 calculates the interference data addition probability according to the following formula:

[0075]

[0076] where P dist is the addition probability of the interference data; q is the number of times the interference data is added in the r-th training cycle of the vaginal grayscale image dataset; Q is the total number of training cycles; D is the number of interference data.

[0077] In a specific embodiment of the present invention, the initial value of q is 5 and gradually decreases to 1 as the training progresses; Q is set to 100; D is set to 500. The interference data addition subunit 2221 randomly selects some data in the vaginal grayscale image dataset according to the calculated addition probability and adds the interference data to the vaginal grayscale image dataset. This interference data addition mechanism can effectively improve the robustness and generalization ability of the model.

[0078] The training processing subunit 2222 is used to train the vaginal grayscale image dataset with the added interference data. During the training process, the training processing subunit 2222 uses the cross-entropy loss function and the Adam optimizer, the initial value of the learning rate is 0.001, and it is reduced to 0.1 times of the original every 30 training cycles.

[0079] The third adjustment unit 223 is used to update the network node weight values of the second convolutional neural network after each training cycle. Specifically, the third adjustment unit 223 includes a weight update subunit 2231 and a bias update subunit 2232.

[0080] The weight update subunit 2231 updates the neural weight matrix of the convolutional neural network according to the following formula:

[0081] W r = W r-1 + η r · M r · E r ,

[0082] where W r is the neural weight matrix of the r-th training cycle; W r-1 is the neural weight matrix of the (r - 1)-th training cycle; η r is the parameter step size of the r-th training cycle; M r is the intermediate matrix of the r-th training cycle; E r is the error matrix of the r-th training cycle.

[0083] The bias update subunit 2232 updates the bias vector of the convolutional neural network according to the following formula:

[0084] B r = B r-1 + η r · V r · E r ,

[0085] where B r is the bias vector of the r-th training cycle; B r -1 is the bias vector of the (r - 1)-th training cycle; η r is the parameter step size of the r-th training cycle; V r is the intermediate vector of the r-th training cycle; E r is the error vector of the r-th training cycle.

[0086] In a specific embodiment of the present invention, the initial value of the parameter step size η r is set to 0.01 and gradually decreases to 0.001 as the training progresses. This parameter update mechanism can ensure that the training process of the neural network can converge quickly and avoid falling into local optimal solutions.

[0087] The feature recognition module 3 is communicatively connected to the neural network processing module 2 and is configured to input the vaginal image to be predicted into the feature classification model set to obtain a feature classification result. As Figure 3 shown, the feature recognition module 3 includes a first calculation unit 31, a second calculation unit 32, and a feature fusion unit 33.

[0088] The first calculation unit 31 is configured to calculate the first confidence level of the vaginal ultrasound image feature classification result. Specifically, the first calculation unit 31 calculates the first confidence level according to the following formula:

[0089]

[0090] Among them, C1 is the first confidence level; L us is the classification loss value of the feature classification model for the vaginal ultrasound image; L gray is the classification loss value of the feature classification model for the vaginal grayscale image.

[0091] The second calculation unit 32 is used to calculate the second confidence level of the feature classification result of the vaginal grayscale image. Specifically, the second calculation unit 32 calculates the second confidence level according to the following formula:

[0092]

[0093] Among them, C2 is the second confidence level; L us is the classification loss value of the feature classification model for the vaginal ultrasound image; L gray is the classification loss value of the feature classification model for the vaginal grayscale image.

[0094] The feature fusion unit 33 is used to perform weighted fusion on the feature classification results of the vaginal ultrasound image and the vaginal grayscale image based on the first confidence level and the second confidence level to obtain the final feature classification result. Specifically, the feature fusion unit 33 calculates the weighted fusion result according to the following formula:

[0095]

[0096] Among them, R is the final feature classification result; R us is the feature classification result of the vaginal ultrasound image; R gray is the feature classification result of the vaginal grayscale image.

[0097] In a specific embodiment of the present invention, when the ratio of the first confidence level C1 to the second confidence level C2 exceeds 3:1 or is less than 1:3, the system will automatically adjust the weight distribution to ensure the reliability of the result. This confidence-based weighted fusion mechanism can effectively improve the accuracy and reliability of feature classification.

[0098] The bed body control module 4 is communicatively connected to the feature recognition module 3 and is used to control the position and posture of the gynecological operating bed based on the feature classification result. As Figure 4 shown, the bed body control module 4 includes a first adjustment unit 41, a second adjustment unit 42, and a third adjustment unit 43.

[0099] The first adjustment unit 41 is used to receive the adjustment instructions for the position of the bed body and control the horizontal and vertical positions of the gynecological operating bed body. Specifically, the first adjustment unit 41 includes a first adjustment motor 411, a rotating rod 412, and a fixed rod 413. The first adjustment motor 411 provides driving force; the rotating rod 412 is connected to the first adjustment motor 411, receives the driving force of the first adjustment motor 411 and rotates; the fixed rod 413 is arranged at the bottom of the operating bed body and is connected to the rotating rod 412, and drives the operating bed body to move under the drive of the rotating rod 412, realizing the adjustment of the horizontal and vertical positions of the operating bed body.

[0100] In a specific embodiment of the present invention, the first adjustment motor 411 is a stepper motor, with a precision of 0.01 mm / step, a maximum rotational speed of 3000 rpm, and a torque of 5 N·m. This configuration can ensure the accuracy and response speed of the bed body position adjustment.

[0101] The second adjustment unit 42 is used to receive the adjustment instructions for the angle of the bed body and control the angle of the gynecological operating bed body. Specifically, the second adjustment unit 42 includes a second adjustment motor 421 and a lead screw 422. The number of the second adjustment units 42 is two, and the two second adjustment units 42 are respectively installed on both sides of the first adjustment unit 41 and perform adjustments synchronously when responding to the adjustment instructions.

[0102] In a specific embodiment of the present invention, the second adjustment motor 421 is also a stepper motor, with a precision of 0.01 degree / step, a maximum rotational speed of 2000 rpm, and a torque of 3 N·m. The lead screw 422 adopts a precision ball screw, with a lead of 5 mm and an accuracy grade of P1. This configuration can ensure the accuracy and stability of the bed body angle adjustment.

[0103] The third adjustment unit 43 is used to receive the adjustment instructions for the height of the bed body and control the height of the gynecological operating bed body. Specifically, the third adjustment unit 43 includes a third adjustment motor 431, a sliding rod 432, and a telescopic rod 433. The sliding rod 432 is arranged inside the telescopic rod 433, and the third adjustment motor 431 drives the sliding rod 432 to slide inside the telescopic rod 433, realizing the extension and retraction of the telescopic rod 433, thereby adjusting the height of the bed body.

[0104] In a specific embodiment of the present invention, the third adjustment motor 431 is a servo motor, with a precision of 0.005 mm / step and a maximum stroke of 500 mm. This configuration can ensure the accuracy and stability of the bed body height adjustment.

[0105] In a preferred embodiment of the present invention, it further includes a remote central control module 5, which is communicatively connected to the bed body control module 4. As Figure 5 shown, the remote central control module 5 includes an image processing unit 51, a command distribution unit 52, and an operation feedback unit 53.

[0106] The image processing unit 51 is used to process the image information from the image acquisition device. Specifically, the image processing unit 51 first preprocesses the acquired image, including operations such as noise removal, contrast enhancement, and edge detection; then performs feature extraction to identify the key information in the image; and finally conducts image analysis to determine the adjustment direction and amplitude required for the bed body.

[0107] In a specific embodiment of the present invention, the image processing unit 51 adopts an object detection algorithm based on a convolutional neural network, such as Faster R-CNN or YOLOv4, to detect the surgical area and key anatomical structures in the image in real time, with an accuracy of up to 98% and a response time of less than 100 ms.

[0108] The command distribution unit 52 is used to generate control commands based on the image information and send them to the bed control module 4. Specifically, the command distribution unit 52 generates specific control commands, such as adjustment instructions for the bed body position, angle, and height, according to the image processing results, in combination with preset surgical procedures and safety parameters, and sends them to the bed control module 4 through the communication interface.

[0109] In a specific embodiment of the present invention, the command distribution unit 52 adopts a combination of a rule-based decision-making system and an adaptive control algorithm based on deep reinforcement learning, which not only ensures the safety of control but also improves the adaptability and intelligence level of the system.

[0110] The operation feedback unit 53 is used to receive the execution information of the bed control module 4 and conduct evaluation processing. Specifically, the operation feedback unit 53 receives the execution status information from the bed control module 4, including the execution completion degree, execution accuracy, and execution time, etc., and compares it with the expected result to generate an evaluation report, providing a reference for subsequent operations.

[0111] In another preferred embodiment of the present invention, an auxiliary system 6 is further included, which is communicatively connected to the remote central control module 5. As Figure 6 shown, the auxiliary system 6 includes an auxiliary bed 61, a positioning module 62, and a position relationship determination unit 63.

[0112] The auxiliary bed 61 serves as an auxiliary tool for the operating bed and can be used to place medical devices, support specific parts of the patient, or assist the doctor in performing specific operations.

[0113] The positioning module 62 is arranged on the auxiliary bed 61 and is used to obtain the position information of the auxiliary bed 61. Specifically, the positioning module 62 can be a GPS module or other position sensors to collect the spatial coordinate information of the auxiliary bed 61 in real time.

[0114] In a specific embodiment of the present invention, the positioning module 62 adopts a high-precision indoor positioning system, such as ultra-wideband (UWB) positioning technology, with a positioning accuracy of up to ±5 cm and an update frequency of 10 Hz. This configuration can ensure the accuracy and real-time nature of the auxiliary bed position information.

[0115] The position relationship determination unit 63 is arranged in the remote central control module 5 and is used to determine the position relationship between the auxiliary bed 61 and the operating bed based on the position information. Specifically, the position relationship determination unit 63 receives the position information from the positioning module 62, combines it with the position information of the operating bed, calculates the relative position and distance between the two, and determines whether it is necessary to adjust the position of the auxiliary bed 61 or the operating bed according to the preset safety parameters and operation requirements.

[0116] In another preferred embodiment of the present invention, the remote central control module 5 further includes a voice recognition unit 54, a voice spatial positioning unit 55, and a voice command processing unit 56.

[0117] The voice recognition unit 54 is used to recognize the content of the voice information. Specifically, the voice recognition unit 54 receives the voice signal from the microphone array, performs noise reduction and enhancement processing, and then uses a voice recognition algorithm to convert the voice signal into text information to recognize the specific content of the voice command.

[0118] In a specific embodiment of the present invention, the voice recognition unit 54 adopts a voice recognition algorithm based on deep learning, such as Deep Speech or Wav2Vec2.0, with a recognition accuracy of up to 95%, supports multiple languages and dialects, and can maintain a high recognition performance in a noisy environment.

[0119] The voice spatial positioning unit 55 is used to determine the position of the voice information sender. Specifically, the voice spatial positioning unit 55 uses the multiple voice signals received by the microphone array and calculates the direction and distance of the sound source through a sound source localization algorithm, thereby determining the spatial position of the voice information sender.

[0120] In a specific embodiment of the present invention, the voice spatial positioning unit 55 adopts a sound source localization algorithm based on time difference of arrival (TDOA), combined with beamforming technology, with a positioning accuracy of up to ±15 degrees and a distance range of 0.5 m to 5 m. This configuration can accurately recognize the voice commands of doctors in different positions in the operating room.

[0121] The voice command processing unit 56 is used to convert voice information into control commands and send them to the bed body control module 4 to control the position or attitude of the operating bed. Specifically, the voice command processing unit 56 generates corresponding control commands, such as "raise", "lower", "turn left", "turn right", etc., according to the recognized voice content and the position of the sender, and combines the preset command mapping rules, and converts these commands into control signals recognizable by the bed body control module 4.

[0122] In yet another preferred embodiment of the present invention, the remote central control module 5 further includes a storage unit 57 and an evaluation processing unit 58.

[0123] The storage unit 57 is used to store surgical information, surgical video information, surgical operation sequence information, and video monitoring information. Specifically, the storage unit 57 uses a large-capacity solid-state drive or network storage device to record various types of information during the surgical process in real time, including changes in the bed body position, image data, operation instructions, and execution status, etc., providing data support for postoperative analysis and quality control.

[0124] In a specific embodiment of the present invention, the storage unit 57 adopts a RAID5 disk array with a total capacity of 10TB, supports hot plugging and data redundancy, ensuring the security and reliability of data.

[0125] The evaluation processing unit 58 is used to perform evaluation processing on surgical operations based on surgical information and surgical video information. Specifically, the evaluation processing unit 58 uses the stored surgical data, combines the preset evaluation criteria and expert experience, comprehensively evaluates the quality, efficiency, and safety of surgical operations, generates an evaluation report, and provides feedback and improvement suggestions for doctors.

[0126] In a specific embodiment of the present invention, the evaluation processing unit 58 adopts a surgical evaluation algorithm based on deep learning, combines surgical specifications and an expert knowledge base, performs multi-dimensional evaluation on surgical operations, including aspects such as operation accuracy, time efficiency, resource utilization, and safety parameters, and the consistency of the evaluation results can reach more than 90%.

[0127] The working process of the present invention will be further described below in combination with specific application scenarios:

[0128] In a typical gynecological operation, first, an intelligent gynecological operating bed control system of the present invention is started. The data acquisition module 1 respectively acquires ultrasonic images and grayscale images of the patient's vaginal area through the ultrasonic image acquisition unit 11 and the grayscale image acquisition unit 12. These images are preprocessed and then input into the feature recognition module 3.

[0129] The feature recognition module 3 inputs the images into the trained vaginal ultrasound image feature classification model and vaginal grayscale image feature classification model respectively, and obtains the preliminary feature classification results. Then, the first calculation unit 31 and the second calculation unit 32 calculate the first confidence level and the second confidence level respectively, and the feature fusion unit 33 performs weighted fusion on the classification results based on these two confidence levels to obtain the final feature classification result.

[0130] The bed body control module 4 automatically generates bed body adjustment instructions according to the feature classification results. The first adjustment unit 41 is responsible for adjusting the horizontal position and vertical position of the bed body, the second adjustment unit 42 is responsible for adjusting the angle of the bed body, and the third adjustment unit 43 is responsible for adjusting the height of the bed body. These adjustment units work together to make the bed body reach the position and posture most suitable for the current surgical state.

[0131] Meanwhile, the remote central control module 5 processes the image information of the surgical area in real time through the image processing unit 51, the command distribution unit 52 generates control commands according to the image processing results, and the operation feedback unit 53 receives the execution information of the bed body control module 4 and conducts evaluation processing. The doctor can also control the system through voice commands. The voice recognition unit 54 recognizes the voice content, the voice spatial positioning unit 55 determines the position of the voice emitter, and the voice command processing unit 56 converts the voice commands into control commands.

[0132] If an auxiliary bed 61 is needed for the operation, the positioning module 62 will obtain the position information of the auxiliary bed 61 in real time, and the position relationship determination unit 63 ensures that the auxiliary bed 61 maintains an appropriate position relationship with the operating bed.

[0133] During the whole operation process, the storage unit 57 records all relevant information, and the evaluation processing unit 58 conducts real-time evaluation on the operation to provide necessary feedback and suggestions for the doctor.

[0134] An intelligent gynecological operating bed control system of the present invention integrates deep learning technology, image processing technology and precision mechanical control technology, realizes intelligent bed body control based on the physiological characteristics of patients, greatly improves the accuracy and safety of gynecological operations, reduces the operation burden of doctors, and provides technical support for telemedicine. In particular, the dual-modal image analysis system and the confidence-weighted feature fusion mechanism significantly improve the accuracy and reliability of feature recognition; while the multi-degree-of-freedom bed body adjustment system realizes precise control of the position and posture of the bed body.

[0135] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. The technical features described in the present invention can be implemented in other specific forms without departing from its spirit or basic characteristics. Therefore, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention all fall within the protection scope of the technical solution of the present invention.

Claims

1. An intelligent control system for a gynecological operating table, characterized in that, Including: A data acquisition module, configured to acquire vaginal ultrasound images and vaginal grayscale images; A neural network processing module, communicatively connected to the data acquisition module, configured to receive the vaginal ultrasound images and the vaginal grayscale images, and train a feature classification model set based on the vaginal ultrasound images and the vaginal grayscale images; A feature recognition module, communicatively connected to the neural network processing module, configured to input a vaginal image to be predicted into the feature classification model set to obtain a feature classification result; A bed body control module, communicatively connected to the feature recognition module, configured to control the position and posture of a gynecological operating bed based on the feature classification result.

2. The intelligent gynecological operating bed control system according to claim 1, characterized in that The neural network processing module includes: A first network construction unit, configured to train the vaginal ultrasound images using a first convolutional neural network to obtain a vaginal ultrasound image feature classification model; A second network construction unit, configured to train the vaginal grayscale images using a second convolutional neural network to obtain a vaginal grayscale image feature classification model; Wherein, the feature classification model set includes the vaginal ultrasound image feature classification model and the vaginal grayscale image feature classification model.

3. The intelligent gynecological operating table control system according to claim 2, wherein The second network construction unit includes: A first adjustment unit, configured to randomly adjust the input order of the vaginal grayscale image data set; A second adjustment unit, configured to add interference data to the vaginal grayscale image data set during the training process; A third adjustment unit, configured to update the network node weight values of the second convolutional neural network after each training cycle ends.

4. The intelligent gynecological operating table control system according to claim 1, characterized in that, The feature recognition module includes: A first calculation unit, configured to calculate a first confidence level of the vaginal ultrasound image feature classification result; A second calculation unit, configured to calculate a second confidence level of the vaginal grayscale image feature classification result; A feature fusion unit, configured to perform weighted fusion on the vaginal ultrasound image feature classification result and the vaginal grayscale image feature classification result based on the first confidence level and the second confidence level to obtain a final feature classification result.

5. The intelligent gynecological operating bed control system according to claim 1, characterized in that, The bed body control module includes: A first adjustment unit, configured to receive an adjustment instruction for the bed body position and control the horizontal position and vertical position of the gynecological operating bed body; A second adjustment unit, configured to receive an adjustment instruction for the bed body angle and control the angle of the gynecological operating bed body; A third adjustment unit, configured to receive an adjustment instruction for the bed body height and control the height of the gynecological operating bed body.

6. The intelligent gynecological operating table control system according to claim 5, characterized in that, The first adjustment unit includes: A first adjustment motor, configured to provide driving force; A rotating rod, connected to the first adjustment motor, configured to receive the driving force of the first adjustment motor and rotate; A fixed rod, disposed at the bottom of the operating bed body, connected to the rotating rod, configured to drive the operating bed body to move under the drive of the rotating rod to realize the adjustment of the horizontal position and vertical position of the operating bed body.

7. The intelligent gynecological operating table control system according to claim 1, wherein It further includes: A remote central control module, communicatively connected to the bed body control module, including: An image processing unit, configured to process image information from an image acquisition device; A command distribution unit, configured to generate a control command based on the image information and send it to the bed body control module; An operation feedback unit, configured to receive the execution information of the bed body control module and perform evaluation processing.

8. An intelligent gynecological operating bed control system according to claim 7, characterized in that, It further includes: The auxiliary system, which is communicatively connected to the remote central control module, includes: An auxiliary bed, which serves as an auxiliary tool for the operating table; A positioning module, which is arranged on the auxiliary bed and is used to obtain the position information of the auxiliary bed; A position relationship determination unit, which is arranged in the remote central control module and is used to determine the position relationship between the auxiliary bed and the operating table based on the position information.

9. The intelligent gynecological operating table control system according to claim 7, wherein, The remote central control module further includes: A voice recognition unit, which is used to recognize the content of voice information; A voice spatial positioning unit, which is used to determine the position of the voice information sender; A voice command processing unit, which is used to convert the voice information into a control command and send it to the bed body control module to control the position or posture of the operating table.

10. The intelligent gynecological operating table control system according to claim 7, characterized in that, The remote central control module further includes: A storage unit, which is used to store surgical information, surgical video information, surgical operation sequence information, and video monitoring information; An evaluation processing unit, which is used to perform evaluation processing on surgical operations based on the surgical information and the surgical video information.

Images