An ultrasonic knife temperature monitoring regulation method, a computer device and a storage medium

Abstract

The application provides an ultrasonic knife temperature monitoring regulation method, computer equipment and a storage medium, and relates to the technical field of medical instruments; the ultrasonic knife temperature monitoring regulation method comprises the following steps: obtaining relevant variables of the temperature of an ultrasonic knife, wherein the relevant variables comprise the contact condition of the head of the ultrasonic knife and body fluid; obtaining target detection boxes corresponding to the body fluid, the ultrasonic knife and the head in a real-time detection image according to a real-time detection image of the ultrasonic knife; predicting the temperature of the ultrasonic knife according to the relevant variables and the target detection boxes corresponding to the body fluid, the ultrasonic knife and the head, and obtaining the temperature change of the ultrasonic knife; and controlling the temperature of the ultrasonic knife through a preset operation according to the temperature change of the ultrasonic knife. The application can more accurately monitor and control the temperature of the ultrasonic knife, improve the controllability and accuracy of the head temperature in the operation process, and improve the temperature monitoring and regulation effect.

Description

Technical Field

[0001] This invention relates to the field of medical device technology, and more specifically, to an ultrasonic scalpel temperature monitoring and control method, a computer device, and a storage medium. Background Technology

[0002] The ultrasonic scalpel, as an innovative surgical cutting tool, has been widely used in the medical field. Utilizing high-frequency ultrasonic vibrations, it enables accurate and rapid tissue cutting while also providing hemostasis. During surgery, the temperature of the ultrasonic scalpel tip needs to be monitored and controlled in real time. This is typically achieved using a traditional temperature sensor to monitor the tip temperature, and then adjusting the temperature accordingly.

[0003] In existing technologies, limitations in the size and location of temperature sensors make it difficult to accurately monitor the actual temperature of the cutting head in real time, leading to inaccurate temperature control. Furthermore, ultrasonic scalpels involve multiple variables related to temperature control, such as cutting head material, tissue type, cutting speed, and the contact between body fluids and the cutting head. However, the influence of these variables on temperature control is often ignored, thus limiting the accuracy and robustness of temperature control. Summary of the Invention

[0004] The problem addressed by this invention is how to improve the effectiveness of temperature monitoring and control.

[0005] To address the above problems, this invention provides an ultrasonic scalpel temperature monitoring and control method, a computer device, and a storage medium.

[0006] In a first aspect, the present invention provides a method for monitoring and controlling the temperature of an ultrasonic scalpel, comprising:

[0007] Acquire relevant variables of the temperature of the ultrasonic scalpel, wherein the relevant variables include the contact between the ultrasonic scalpel tip and the body fluid;

[0008] Based on the real-time detection image of the ultrasonic scalpel, target detection boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head are obtained in the real-time detection image, respectively.

[0009] Based on the relevant variables and the target detection boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head, the temperature of the ultrasonic scalpel is predicted, and the temperature change of the ultrasonic scalpel is obtained.

[0010] The temperature of the ultrasonic scalpel is controlled by a preset operation based on the temperature change of the ultrasonic scalpel.

[0011] Optionally, the variables related to obtaining the temperature of the ultrasonic scalpel include:

[0012] The real-time detected image is input into the pixel communication model, and the pixel value of each pixel in the target region of the real-time detected image is obtained through the input layer of the pixel communication model.

[0013] Based on the pixel value, the activation value of each pixel is obtained through the fully connected layer of the pixel communication model;

[0014] Based on the activation value and pixel value of each pixel, the output value of each pixel is obtained through the output layer of the pixel communication model;

[0015] The contact status between the blade and the bodily fluid is obtained based on the output value of the pixel in the target area.

[0016] Optionally, determining the contact status between the blade and the bodily fluid based on the output value in the target area includes:

[0017] When the output value of the pixel in the target area continuously increases and remains greater than 0 within a first preset time range, it is determined that the body fluid is in contact with the blade.

[0018] Optionally, obtaining the target detection boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head in the real-time detection image based on the real-time detection image of the ultrasonic scalpel includes:

[0019] Based on the shape characteristics of the body fluid, the ultrasonic scalpel, and the scalpel head, corresponding prior boxes are established respectively;

[0020] Based on the prior boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head, regression prediction boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head in the real-time detection image are obtained respectively.

[0021] The regression prediction boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head are corrected to obtain the target detection boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head, respectively.

[0022] Optionally, the step of correcting the regression prediction boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head respectively to obtain the target detection boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head respectively includes:

[0023] Using the HSV color model, the red pixel value in the regression prediction box of the body fluid is obtained based on the pixel points in the regression prediction box of the body fluid.

[0024] The red pixel value indicates the hue, saturation, and brightness of the HSV color model.

[0025] Based on the red pixel value, the correction judgment factor of the regression prediction box of the body fluid is obtained through the blood regression box correction formula;

[0026] The target detection box of the body fluid is obtained based on the modified judgment factor.

[0027] Optionally, the step of correcting the regression prediction boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head respectively to obtain the target detection boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head respectively further includes:

[0028] Based on the time series formula, the loss value of the regression prediction box of the ultrasonic scalpel and the scalpel head is obtained;

[0029] Based on the loss values ​​of the ultrasonic scalpel and the scalpel head and the surgical decision order, the regression prediction boxes of the ultrasonic scalpel and the scalpel head are corrected to obtain the target detection boxes of the ultrasonic scalpel and the scalpel head.

[0030] Optionally, the step of predicting the temperature of the ultrasonic scalpel based on the relevant variables and the target detection boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head, respectively, to obtain the temperature change of the ultrasonic scalpel, includes:

[0031] Based on the real-time detection images at two time points, the positional change of the target detection box of the ultrasonic scalpel is obtained;

[0032] The cutting speed of the ultrasonic scalpel is obtained based on the position change.

[0033] By normalizing the data, the feature vector at the current moment is obtained based on the cutting speed of the ultrasonic scalpel, the contact between the scalpel tip and the body fluid, and the relevant variables.

[0034] The feature vector at the current moment when the cutting speed exceeds a preset threshold, the body fluid comes into contact with the blade, or the ultrasonic scalpel is in a water-cooled state is encoded and input to obtain a connection vector;

[0035] Based on the connection vector, the temperature change of the ultrasonic scalpel within a second preset time range is obtained.

[0036] Optionally, controlling the temperature of the ultrasonic scalpel according to the temperature change of the ultrasonic scalpel through a preset operation includes:

[0037] Based on the temperature change of the ultrasonic scalpel within the second preset time range, the temperature of the ultrasonic scalpel at any time point within the preset time range is obtained;

[0038] When the temperature at any given time exceeds a safety threshold, the temperature of the ultrasonic scalpel is controlled to decrease to within the safety threshold through a preset operation.

[0039] In a second aspect, the present invention provides a computer device, including a computer-readable storage medium storing a computer program and a processor, wherein the computer program is read and executed by the processor to implement the ultrasonic scalpel temperature monitoring and control method described above.

[0040] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the ultrasonic scalpel temperature monitoring and control method described above.

[0041] The ultrasonic scalpel temperature monitoring and control method, system, and storage medium of this invention acquire relevant variables of the ultrasonic scalpel's temperature, enabling more accurate monitoring of the actual temperature of the scalpel tip and solving the inaccuracy of real-time temperature monitoring. Furthermore, by obtaining target detection boxes corresponding to body fluid, the ultrasonic scalpel, and the scalpel tip in the real-time detection image based on the ultrasonic scalpel's real-time detection image, image processing and computer vision technologies are utilized to analyze target regions in the real-time detection image, achieving the localization and identification of body fluid, the ultrasonic scalpel, and the scalpel tip. By determining the positions of these targets, relevant variables can be more accurately associated with the corresponding target regions. Based on the above steps, the temperature of the ultrasonic scalpel is predicted according to the relevant variables and target detection boxes, yielding temperature changes. Machine learning or deep learning methods are used to more accurately predict temperature change trends, preparing for subsequent temperature control to meet surgical needs. This invention can more accurately monitor and control the temperature of the ultrasonic scalpel, improving the controllability and accuracy of the scalpel tip temperature during surgery, and enhancing the effectiveness of temperature monitoring and control. Attached Figure Description

[0042] Figure 1 This is one of the flowcharts for the ultrasonic scalpel temperature monitoring and control method of the present invention;

[0043] Figure 2 This is the second flowchart of the ultrasonic scalpel temperature monitoring and control method of the present invention;

[0044] Figure 3 This is the third flowchart of the ultrasonic scalpel temperature monitoring and control method of the present invention;

[0045] Figure 4 This is the fourth flowchart of the ultrasonic scalpel temperature monitoring and control method of the present invention;

[0046] Figure 5 This is the fifth flowchart of the ultrasonic scalpel temperature monitoring and control method of the present invention;

[0047] Figure 6This is the sixth flowchart of the ultrasonic scalpel temperature monitoring and control method of the present invention;

[0048] Figure 7 This is the seventh flowchart of the ultrasonic scalpel temperature monitoring and control method of the present invention;

[0049] Figure 8 This is one of the Yolov5 network architecture diagrams of the present invention;

[0050] Figure 9 This is the second diagram of the Yolov5 network architecture of the present invention;

[0051] Figure 10 This is the Seq2Seq network architecture of the present invention. Detailed Implementation

[0052] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0053] Firstly, combining Figure 1 As shown, the present invention provides a method for monitoring and controlling the temperature of an ultrasonic scalpel, comprising:

[0054] The relevant variables of the ultrasonic scalpel temperature are obtained, wherein the relevant variables include the contact between the ultrasonic scalpel tip and the body fluid.

[0055] Specifically, the temperature-related variables of the ultrasonic scalpel are acquired. Changes in these variables cause changes in the temperature of the ultrasonic scalpel tip. These variables include the ultrasonic scalpel's operation time, power setting, tip material, tissue type, tip surface area, cooling mechanism, cutting speed, and whether there is blood contact. In a preferred embodiment of this invention, advanced image vision algorithms are used to determine whether there is blood or body fluid contact. Furthermore, Labelimg is used to annotate the ultrasonic scalpel in the image. Notably, because the ultrasonic scalpel tip is relatively long and narrow, this application uses a separate annotation method for the tip to avoid the situation where only the tip is present in the target detection area, and the ultrasonic scalpel handle is not visible, thus preventing the ultrasonic scalpel from being undetectable.

[0056] Based on the real-time detection image of the ultrasonic scalpel, target detection boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head are obtained in the real-time detection image.

[0057] Specifically, in combination Figure 8 and Figure 9As shown, in a preferred embodiment of the present invention, an ultrasonic scalpel target detection method based on an improved Yolov5 network architecture is employed. This method performs high-order feature extraction on the input real-time detection image and employs a four-layer CBS and C2F architecture. The CBS consists of Conv, BN, and SiLU; the C2F consists of CBS, Split, and Bottleneck; and the Bottleneck consists of CBS and residual blocks. This yields the target detection boxes corresponding to body fluid, the ultrasonic scalpel, and the scalpel head in the real-time detection image, thereby enhancing the accuracy of target detection for the scalpel head, body fluid, and ultrasonic scalpel. In some preferred embodiments, detailed image data related to the surgical ultrasonic scalpel is collected. To make the algorithm model training more efficient, the collected image format is uniformly converted, and the training set image size is standardized. The standardized image size can be 648*648 or another size.

[0058] Based on the relevant variables and the target detection boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head, the temperature of the ultrasonic scalpel is predicted, and the temperature change of the ultrasonic scalpel is obtained.

[0059] Specifically, during different stages of ultrasonic scalpel surgery, the temperature is significantly affected by factors such as whether there is bleeding, the cooling method, and the cutting speed. In order to capture these effects more accurately, this invention uses the Fzclass branch operator for calculation. In addition, a branch network with an encoding end that can input a long sequence vector is used to integrate and analyze the feature parameters of each process of ultrasonic scalpel during the entire surgery. The encoded vector is then input into the connection vector, thereby more sensitively capturing the feature changes of each operation stage and accurately predicting the temperature of ultrasonic scalpel.

[0060] The temperature of the ultrasonic scalpel is controlled by a preset operation based on the temperature change of the ultrasonic scalpel.

[0061] Specifically, by predicting the temperature changes of the scalpel in real time, and through preset operations, it can provide early warnings of potential non-standard operations during surgery. That is, predicting excessively high temperatures strongly suggests that medical personnel may be engaging in non-compliant procedures during surgery, thereby ensuring the proper use of equipment and the standardization of personnel operations throughout the entire surgical process. It can also control the ultrasonic scalpel to regulate its temperature.

[0062] The ultrasonic scalpel temperature monitoring and control method of this invention obtains relevant variables of the ultrasonic scalpel's temperature, enabling more accurate monitoring of the actual temperature of the scalpel tip and solving the inaccuracy of real-time temperature monitoring. Furthermore, based on the real-time detection image of the ultrasonic scalpel, target detection boxes corresponding to the body fluid, ultrasonic scalpel, and scalpel tip are obtained within the real-time detection image. Utilizing image processing and computer vision technology, the target regions in the real-time detection image are analyzed to achieve the localization and identification of the body fluid, ultrasonic scalpel, and scalpel tip. By determining the positions of these targets, relevant variables can be more accurately associated with the corresponding target regions. Based on the above steps, the temperature of the ultrasonic scalpel is predicted according to the relevant variables and target detection boxes, yielding temperature changes. Machine learning or deep learning methods are used to more accurately predict the temperature change trend, preparing for subsequent temperature control to meet surgical needs. This invention can more accurately monitor and control the temperature of the ultrasonic scalpel, improving the controllability and accuracy of the scalpel tip temperature during surgery, and enhancing the effectiveness of temperature monitoring and control.

[0063] Optionally, combined Figure 2 As shown, the relevant variables for obtaining the temperature of the ultrasonic scalpel include the contact between the ultrasonic scalpel tip and the body fluid, including:

[0064] The real-time detected image is input into the pixel communication model, and the pixel value of each pixel in the target region of the real-time detected image is obtained through the input layer of the pixel communication model.

[0065] Based on the pixel value, the activation value of each pixel is obtained through the fully connected layer of the pixel communication model;

[0066] Based on the activation value and pixel value of each pixel, the output value of each pixel is obtained through the output layer of the pixel communication model;

[0067] The contact status between the blade and the bodily fluid is obtained based on the output value of the pixel in the target area.

[0068] Specifically, since the relevant variables include multiple aspects such as the ultrasonic scalpel's operation time, power setting, tip material, tissue type, tip surface area, cooling mechanism, cutting speed, and whether there will be blood contact, it is necessary to acquire each data point individually. The tip material and tip surface area are obtained according to the ultrasonic scalpel's manufacturer's manual parameters. The power and operation time can be obtained through the display on the ultrasonic host. The tissue type can be identified using the YOLO algorithm, including different tissues such as epithelial tissue, connective tissue, and muscle tissue. Blood contact is detected using a pixel transmission model.

[0069] For blood contact determination, this stage uses a pixel-based transmission model to detect the presence of high-value pixels in the target area. That is, bleeding will result in a large amount of bright red color, which represents the presence of high-value pixels within the image. Using the input layer, fully connected layer, and output layer of the pixel-based transmission model, the pixel value, activation value, and output value of each pixel are obtained. The contact status between the blade and the bodily fluid is then determined based on the output value.

[0070] The first input layer uses a convolutional neural network to extract pixel values. The convolutions are in a 3x3 format, with all values ​​being 1 and intervals of 3. The pixel values ​​of each 3x3 region are accumulated to form a single value. Following the second layer is a fully connected layer, with the activation of the second layer as follows:

[0071]

[0072] Where y is the value after activation, x is the value within the neuron, and μ is the average pixel value of the image at this stage when there is no bleeding. This layer allows the value to be larger at the location of bleeding, and the value to be biased towards 0 if there is no bleeding.

[0073] In this embodiment, by predicting and analyzing the pixel value of each pixel, a detailed judgment can be made as to whether bodily fluids such as blood are in contact with the blade.

[0074] Optionally, determining the contact status between the blade and the bodily fluid based on the output value in the target area includes:

[0075] When the output value of the pixel in the target area continuously increases and remains greater than 0 within a first preset time range, it is determined that the body fluid is in contact with the blade.

[0076] Specifically, through this network, the algorithm model can make a more comprehensive judgment on whether there is bleeding during the use of the ultrasonic scalpel. When the output value fluctuates, that is, it is continuously greater than 0 and continues to increase, a judgment can be made directly to determine that the body fluid has come into contact with the scalpel head, that is, bleeding has occurred.

[0077] In this embodiment, by predicting and analyzing the pixel value of each pixel, a detailed judgment can be made as to whether bodily fluids such as blood are in contact with the blade.

[0078] Optionally, combined Figure 3 As shown, the step of obtaining target detection boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head in the real-time detection image based on the real-time detection image of the ultrasonic scalpel includes:

[0079] Based on the shape characteristics of the body fluid, the ultrasonic scalpel, and the scalpel head, corresponding prior boxes are established respectively;

[0080] Based on the prior boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head, regression prediction boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head in the real-time detection image are obtained respectively.

[0081] The regression prediction boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head are corrected to obtain the target detection boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head, respectively.

[0082] Specifically, prior boxes are first established based on the shape characteristics of the body fluid, ultrasonic scalpel, and scalpel head. In a preferred embodiment of the invention, the prior boxes for the scalpel head are [10,40], [16,52], and [25,100], the prior boxes for blood are [5,5], [10,10], and [20,20], and the prior boxes for the ultrasonic scalpel are [10,13], [16,30], and [20,40]. This application employs customized prior boxes for different target objects. These prior boxes are designed according to the actual shape characteristics of the objects. The scalpel head uses a long and narrow box, while the blood and body fluids use a smaller square box. This targeted design allows the prior boxes to better fit the actual shape of the target object, improving the model's positioning accuracy against the scalpel head and blood clots. Convolution calculations, i.e., C2F and CBS calculations, are then performed on different network layers to output regression prediction values ​​and classification prediction values ​​for each target.

[0083] The regression prediction frames are then corrected and validated to obtain the target detection frames for body fluids, ultrasonic scalpels, and scalpel heads.

[0084] In this embodiment, by using customized prior boxes for different target objects, the prior boxes can better fit the actual shape of the target objects, thereby improving the positioning accuracy of the model against the head and blood clots.

[0085] Optionally, combined Figure 4 As shown, the step of correcting the regression prediction boxes for the body fluid, the ultrasonic scalpel, and the scalpel head to obtain the target detection boxes for the body fluid, the ultrasonic scalpel, and the scalpel head includes:

[0086] Using the HSV color model, the red pixel value in the regression prediction box of the body fluid is obtained based on the pixel points in the regression prediction box of the body fluid.

[0087] The red pixel value indicates the hue, saturation, and brightness of the HSV color model.

[0088] Based on the red pixel value, the correction judgment factor of the regression prediction box of the body fluid is obtained through the blood regression box correction formula;

[0089] The target detection box of the body fluid is obtained based on the modified judgment factor.

[0090] Specifically, the blood regression frame is corrected according to the blood regression formula.

[0091] The blood regression box correction formula is as follows:

[0092]

[0093]

[0094]

[0095] los=1-σ(4)

[0096] Where M is the ROI region of the blood regression box in the original image, M(x,y) represents the pixel value of the image at coordinates (x,y), the size of the regression box in the original image is W*P, the width is W, the height is P, H, S, and V represent the hue, saturation, and brightness in the HSV color model, respectively, σ is the correction factor, and los is the HSV loss value.

[0097] The blood regression box can be corrected using the above formulas (1), (2), (3), and (4). First, the pixel values ​​in the blood regression box are filtered using HSV to extract the red pixel values. Only when the correction judgment factor is greater than 0.8 will the regression box be judged as correctly classified. In addition, this application uses the HSV loss value as the loss of the blood detection regression box in the improved Yolov5 network architecture. In the HSV color model, hue is H, saturation is S, and brightness is V. Therefore, the red pixel values ​​extracted according to the HSV color model are: hue H is in the range close to red (e.g., 0 to 10 or 170 to 180), saturation S needs to be greater than a certain threshold (representing the strength of color saturation), and brightness V needs to be greater than a certain threshold (representing color brightness) to be considered as red pixel values.

[0098] In this embodiment, a blood bounding box correction technique is introduced for ultrasonic scalpel target detection to improve the recognition effect of blood and enhance the performance, reliability and robustness of the system.

[0099] Optionally, combined Figure 5 As shown, the step of correcting the regression prediction boxes for the body fluid, the ultrasonic scalpel, and the scalpel head to obtain the target detection boxes for the body fluid, the ultrasonic scalpel, and the scalpel head includes:

[0100] Based on the time series formula, the loss value of the regression prediction box of the ultrasonic scalpel and the scalpel head is obtained;

[0101] Based on the loss values ​​of the ultrasonic scalpel and the scalpel head and the surgical decision order, the regression prediction boxes of the ultrasonic scalpel and the scalpel head are corrected to obtain the target detection boxes of the ultrasonic scalpel and the scalpel head.

[0102] Specifically, firstly, a large number of images of ultrasonic scalpels and surgical tips in real surgical scenarios are collected, and target detection boxes are labeled for each image. The labeling process includes information related to loss values ​​and surgical decision order. Then, using the collected training data, a suitable model for target detection is selected and trained. During training, loss values ​​and surgical decision order can be introduced as additional supervision information to help the model learn regression prediction boxes for the ultrasonic scalpel and surgical tip. In application, for new surgical scene images, the trained target detection model is used to predict the regression prediction boxes for the ultrasonic scalpel and surgical tip; simultaneously, the loss value is calculated based on the position and shape of the regression prediction boxes. Then, based on the loss value and surgical decision order information, the degree of correction for the regression prediction boxes can be determined. Using the loss value and surgical decision order as parameters, and based on empirical rules or machine learning methods, the regression boxes are corrected to obtain more accurate target detection boxes.

[0103] This application proposes a method for determining the regression results of the scalpel and ultrasonic scalpel based on surgical time series, and assigns the loss according to a time series formula.

[0104] The timing formula is as follows:

[0105] L total =L basic -ω*lnP(A / B) (5);

[0106] Among them, L total L represents the total loss value of the regression box. basic The target detection loss is based on ω, where ω is the weighting coefficient and P(A / B) is the probability of the ultrasonic scalpel appearing at that time point.

[0107] The above formula can be used to correct the return of the blade and ultrasonic scalpel to the frame according to the surgical judgment sequence.

[0108] In this embodiment, the correction of the blade and ultrasonic scalpel return frame is achieved based on the results of the surgical time series, thereby improving the system's performance, reliability, and robustness.

[0109] Optionally, combined Figure 6As shown, the step of predicting the temperature of the ultrasonic scalpel based on the relevant variables and the target detection boxes corresponding to the body fluid, the ultrasonic scalpel, and the scalpel head, respectively, to obtain the temperature change of the ultrasonic scalpel, includes:

[0110] Based on the real-time detection images at two time points, the positional change of the target detection box of the ultrasonic scalpel is obtained;

[0111] The cutting speed of the ultrasonic scalpel is obtained based on the position change.

[0112] By normalizing the data, the feature vector at the current moment is obtained based on the cutting speed of the ultrasonic scalpel, the contact between the scalpel tip and the body fluid, and the relevant variables.

[0113] The feature vector at the current moment when the cutting speed exceeds a preset threshold, the body fluid comes into contact with the blade, or the ultrasonic scalpel is in a water-cooled state is encoded and input to obtain a connection vector;

[0114] Based on the connection vector, the temperature change of the ultrasonic scalpel within a second preset time range is obtained.

[0115] Specifically, firstly, this application uses the cutting speed of the ultrasonic scalpel to determine the cutting speed, which can be done using a cutting speed determination formula. The change in the position of the target detection box is used to determine the change in the coordinates of the center point of the target detection box.

[0116] The formula for determining the cutting speed is as follows:

[0117]

[0118] Where v is the speed at which the ultrasonic scalpel is moved, x1 and x2 represent the x-coordinates of the ultrasonic scalpel frame center at two different times, y1 and y2 represent the y-coordinates of the ultrasonic scalpel frame center at two different times, and t is the time it takes for the ultrasonic scalpel to move.

[0119] During the surgery, since most of the cutting actions are only involved in the horizontal cutting of tissues and organs, the cutting speed of the ultrasonic scalpel in this application mainly depends on the speed of movement of the horizontal axis, while the speed of movement of the vertical axis is given a certain weight according to the type of surgery.

[0120] The study identified several variables affecting the temperature of the ultrasonic scalpel tip during surgery, including operation time, power setting, tip material, tissue type, tip surface area, cooling mechanism, cutting speed, and the presence of blood. Tip materials were categorized as stainless steel, ceramic, and titanium alloy; tissue types as epithelial, connective, and muscle tissue; and cooling mechanisms as passive cooling and water cooling systems. The presence of blood contact was indicated by discrete values ​​of 0 and 1.

[0121] The above eigenvalues ​​are normalized and formed into a 1x8 vector, with each vector marked with a time point.

[0122] Combination Figure 10 As shown, this network architecture aims to predict the temporal temperature of ultrasonic scalpels by proposing and implementing an innovative Seq2Seq network architecture. During various stages of ultrasonic scalpel surgery, temperature is significantly affected by factors such as bleeding, cooling methods, and cutting speed. To capture these effects more accurately, this invention employs a branch operator called Fzclass, which can classify a long sequence of 1x8 feature vectors. If, within a certain period, some ultrasonic scalpel operations involve contact between the scalpel tip and blood, or the cooling method is water cooling rather than passive cooling, or rapid cutting occurs, the Fzclass operator will filter the vectors at these specific time points and use them as one of the encoding inputs. This achieves encoding inputs from three branches, thus capturing various changes during the surgical process more meticulously. Finally, these three encoding results are integrated and input into the connection vector C. Furthermore, a branch network with an encoding end capable of inputting a long sequence vector is used to comprehensively analyze the feature parameters of each ultrasonic scalpel process throughout the entire surgical procedure, and the encoded vector is input into the connection vector C. This improved architecture model can more sensitively capture the characteristic changes at each stage of the operation, thereby enhancing the ability to identify and handle abnormal temperature fluctuations. Secondly, the multi-branch input strategy enables the network to gain a deeper understanding of the surgical process from multiple dimensions, further improving the accuracy and robustness of temperature prediction. Finally, integrating multiple encoding results can comprehensively reflect various information during the surgical process, enabling accurate prediction of the ultrasonic scalpel temperature.

[0123] In this embodiment, the cutting speed of the ultrasonic scalpel is obtained through image processing, and then the cutting speed is included in the temperature prediction range, thereby improving the accuracy of ultrasonic scalpel temperature prediction.

[0124] Optionally, combined Figure 7 As shown, controlling the temperature of the ultrasonic scalpel according to the temperature change of the ultrasonic scalpel through a preset operation includes:

[0125] Based on the temperature change of the ultrasonic scalpel within the second preset time range, the temperature of the ultrasonic scalpel at any time point within the preset time range is obtained;

[0126] When the temperature at any given time exceeds a safety threshold, the temperature of the ultrasonic scalpel is controlled to decrease to within the safety threshold through a preset operation.

[0127] Specifically, during the surgery, the real-time temperature prediction system, once it predicts that the temperature is about to exceed a set safety threshold, will, through preset operations, reduce the temperature of the ultrasonic scalpel to within the safety threshold and trigger audible, visual, or vibration alarms to notify the operator. In a preferred embodiment of the invention, the preset operations can be to activate an additional cooling mechanism to rapidly reduce the scalpel tip temperature and simultaneously adjust the power and cutting speed of the ultrasonic scalpel to avoid continuous high-intensity work. If necessary, the surgery will be temporarily interrupted to allow the equipment to cool to a safe temperature before resuming. Furthermore, by predicting the scalpel temperature in real time, it can also provide early warnings of potential non-compliant operations during the surgery; that is, predicting excessively high temperatures may indicate non-compliant operations by medical personnel during the surgery, thereby ensuring the reasonable compliance of equipment and the standardization of personnel operations throughout the entire surgical process.

[0128] In this embodiment, not only can the device temperature be warned in real time, but also non-standard operations by medical staff can be detected and warned through temperature monitoring, which greatly improves the safety and standardization of the surgical procedure.

[0129] In a second aspect, the present invention provides a computer device, including a computer-readable storage medium storing a computer program and a processor, wherein the computer program is read and executed by the processor to implement the ultrasonic scalpel temperature monitoring and control method described above.

[0130] The ultrasonic scalpel temperature monitoring and control system of this invention acquires relevant variables of the ultrasonic scalpel's temperature, enabling more accurate monitoring of the actual temperature of the scalpel tip and solving the inaccuracy of real-time temperature monitoring. Furthermore, based on the real-time detection image of the ultrasonic scalpel, target detection boxes corresponding to the body fluid, ultrasonic scalpel, and scalpel tip are obtained within the real-time detection image. Utilizing image processing and computer vision technologies, the system analyzes target regions in the real-time detection image to achieve the localization and identification of the body fluid, ultrasonic scalpel, and scalpel tip. By determining the positions of these targets, relevant variables can be more accurately associated with their corresponding target regions. Based on the above steps, the temperature of the ultrasonic scalpel is predicted according to the relevant variables and target detection boxes, yielding temperature changes. Machine learning or deep learning methods are used to more accurately predict temperature change trends, preparing for subsequent temperature control to meet surgical needs. This invention can more accurately monitor and control the temperature of the ultrasonic scalpel, improving the controllability and accuracy of the scalpel tip temperature during surgery and enhancing the effectiveness of temperature monitoring and control.

[0131] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the ultrasonic scalpel temperature monitoring and control method described above is implemented.

[0132] The readable storage medium of this invention, by acquiring relevant variables of the ultrasonic scalpel's temperature, more accurately monitors the actual temperature of the scalpel tip, solving the inaccuracy of real-time temperature monitoring. Furthermore, by obtaining target detection boxes corresponding to body fluids, the ultrasonic scalpel, and the scalpel tip in the real-time detection image based on the ultrasonic scalpel's real-time detection image, and utilizing image processing and computer vision technologies, the location and identification of body fluids, the ultrasonic scalpel, and the scalpel tip are achieved by analyzing the target regions in the real-time detection image. By determining the positions of these targets, relevant variables can be more accurately associated with the corresponding target regions. Based on the above steps, the temperature of the ultrasonic scalpel is predicted according to the relevant variables and target detection boxes, obtaining temperature changes. Machine learning or deep learning methods are used to more accurately predict the temperature change trend, preparing for subsequent temperature control to meet surgical needs. This invention can more accurately monitor and control the temperature of the ultrasonic scalpel, improving the controllability and accuracy of the scalpel tip temperature during surgery, and enhancing the effectiveness of temperature monitoring and control.

[0133] While the present invention has been disclosed above, its scope of protection is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention, and all such changes and modifications will fall within the scope of protection of the present invention.

Claims

1. An ultrasonic shears temperature monitoring regulatory system, characterized by, The method comprises the following steps: acquiring the contact condition of the blade head of the ultrasonic knife and the body fluid, specifically comprising: inputting the real-time detection image of the ultrasonic knife into a pixel transmission model, and obtaining the pixel value of each pixel point in the target region of the real-time detection image through the input layer of the pixel transmission model; obtaining the activation value of each pixel point through the full connection layer of the pixel transmission model according to the pixel value; obtaining the output value of each pixel point through the output layer of the pixel transmission model according to the activation value and the pixel value of each pixel point; obtaining the contact condition of the blade head and the body fluid according to the output value of the pixel point in the target region; when the output value of the pixel point in the target region continuously increases and continuously exceeds 0 within a first preset time range, it is determined that the body fluid is in contact with the blade head; obtaining the target detection box corresponding to the body fluid, the ultrasonic knife and the blade head in the real-time detection image of the ultrasonic knife; predicting the temperature of the ultrasonic knife according to the contact condition of the blade head of the ultrasonic knife and the body fluid and the target detection box corresponding to the body fluid, the ultrasonic knife and the blade head, and obtaining the temperature change of the ultrasonic knife; specifically comprising: obtaining the position change of the target detection box of the ultrasonic knife according to the real-time detection images at two time points; obtaining the cutting speed of the ultrasonic knife according to the position change; obtaining the feature vector at the current time through normalization processing according to the cutting speed of the ultrasonic knife, the contact condition of the blade head and the body fluid, and the water cooling state of the ultrasonic knife; encoding and inputting the feature vector at the current time when the cutting speed exceeds the preset threshold or the body fluid is in contact with the blade head or the ultrasonic knife is in the water cooling state to obtain the connection vector; obtaining the temperature change of the ultrasonic knife within a second preset time range according to the connection vector; controlling the temperature of the ultrasonic knife through a preset operation according to the temperature change of the ultrasonic knife.

2. The ultrasonic shears temperature monitoring regulatory system of claim 1, wherein, The method comprises the following steps: establishing the corresponding prior box according to the shape characteristics of the body fluid, the ultrasonic knife and the blade head; obtaining the regression prediction box corresponding to the body fluid, the ultrasonic knife and the blade head in the real-time detection image according to the prior box corresponding to the body fluid, the ultrasonic knife and the blade head; correcting the regression prediction box corresponding to the body fluid, the ultrasonic knife and the blade head to obtain the target detection box corresponding to the body fluid, the ultrasonic knife and the blade head.

3. The ultrasonic shear temperature monitoring regulatory system of claim 2, wherein, The method comprises the following steps: According to the pixel points in the regression prediction frame of the body fluid, a red pixel value in the regression prediction frame of the body fluid is obtained through an HSV color model; The red pixel value indicates hue, saturation, and brightness of the HSV color model; According to the red pixel value, a correction judgment factor of the regression prediction frame of the body fluid is obtained through a blood regression frame correction formula; According to the correction judgment factor, the target detection frame of the body fluid is obtained.

4. The ultrasonic shear temperature monitoring regulatory system of claim 3, wherein, The method further includes: According to a time sequence formula, a loss value of the regression prediction frame of the ultrasonic knife and the blade head is obtained; According to the loss value of the ultrasonic knife and the blade head and a surgical judgment sequence, the correction of the regression prediction frame of the ultrasonic knife and the blade head is performed to obtain the target detection frame of the ultrasonic knife and the blade head.

5. The ultrasonic shear temperature monitoring regulatory system of claim 4, wherein, The method further includes: According to the temperature change of the ultrasonic knife within the second preset time range, a temperature of the ultrasonic knife at an arbitrary time point within the preset time range is obtained; when the temperature at the arbitrary time point exceeds a safety threshold, the temperature of the ultrasonic knife is controlled to decrease to within the safety threshold through a preset operation.

Images