Methods, devices, computer equipment, and storage media for predicting the timing of puncture.

Abstract

This application relates to a method, apparatus, computer device, storage medium, and computer program product for predicting puncture timing. The method includes: acquiring the actual respiratory waveforms of the subject for the preceding several cycles in the next cycle; inputting the actual respiratory waveforms of the preceding cycles into a pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform for the next cycle; acquiring the target respiratory feature value of the subject; the target respiratory feature value is the respiratory feature value of the subject when acquiring the medical scan image used for needle tract planning; and determining the predicted respiratory feature value that matches the target respiratory feature value in the predicted respiratory waveform for the next cycle, thereby determining the puncture timing in the next cycle. This method can improve the accuracy of puncture timing prediction.

Description

Technical Field

[0001] This application relates to the field of medical technology, and in particular to a method, apparatus, computer device, storage medium, and computer program product for predicting puncture timing. Background Technology

[0002] Interventional surgery refers to the introduction of specialized catheters, guidewires, puncture needles, and other precision instruments into the human body under the guidance of medical imaging equipment to diagnose and treat internal diseases. Interventional therapy is characterized by being non-surgical, minimally invasive, having a rapid recovery, and yielding good results.

[0003] In actual interventional surgery, doctors use medical imaging to determine the lesion area, then identify the target point, and determine the final puncture timing based on experience to deliver the puncture instrument to the target point for surgical treatment. However, under anesthesia, the patient's breathing or other physiological activities can cause the target point to deviate from the actual target point planned in the imaging. Alternatively, due to the time lag between image imaging and the final execution by the robotic arm, the actual target point reached may differ from the theoretical target point. In other words, the judgment of puncture timing in related technologies cannot match the actual situation, resulting in the inability to accurately deliver the puncture instrument to the target point.

[0004] Therefore, there is a problem with inaccurate prediction of puncture timing in related technologies during interventional procedures. Summary of the Invention

[0005] Therefore, it is necessary to provide a method, device, computer equipment, computer-readable storage medium, and computer program product for predicting puncture timing that can improve the accuracy of puncture timing prediction during interventional surgery, in order to address the above-mentioned technical problems.

[0006] Firstly, this application provides a method for predicting the timing of puncture. The method includes:

[0007] For the upcoming next cycle, obtain the actual respiratory waveform of the punctured subject in the previous several cycles of the next cycle;

[0008] The actual respiratory waveforms corresponding to the previous several cycles are input into the pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform for the next cycle.

[0009] Obtain the target respiratory feature value of the punctured object; the target respiratory feature value is the respiratory feature value of the punctured object when acquiring the medical scan image used for needle path planning of the puncture needle;

[0010] In the predicted respiratory waveform corresponding to the next cycle, a predicted respiratory feature value that matches the target respiratory feature value is determined in order to determine the timing of puncture in the next cycle.

[0011] In one embodiment, the step of inputting the actual respiratory waveforms corresponding to the previous several cycles into a pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform for the next cycle includes:

[0012] Discretize the continuous actual respiratory waveforms corresponding to the first few cycles to obtain the discrete actual respiratory waveforms corresponding to the first few cycles.

[0013] The discrete actual respiratory waveforms corresponding to the previous several cycles are input into the pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform for the next cycle.

[0014] In one embodiment, discretizing the continuous actual respiratory waveform corresponding to the first few cycles to obtain the discrete actual respiratory waveform corresponding to the first few cycles includes:

[0015] Based on the actual respiratory characteristic values ​​corresponding to the continuous actual respiratory waveforms of the previous several cycles at discrete time points, the discrete actual respiratory waveforms corresponding to the previous several cycles are obtained.

[0016] In one embodiment, the step of inputting the discrete actual respiratory waveforms corresponding to the previous several cycles into the pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform for the next cycle includes:

[0017] The discrete actual respiratory waveforms corresponding to the previous several cycles are input into the pre-trained respiratory waveform prediction model to obtain the discrete predicted respiratory waveforms corresponding to the next cycle.

[0018] The discrete predicted respiratory waveform corresponding to the next cycle is interpolated to obtain the predicted respiratory waveform corresponding to the next cycle.

[0019] In one embodiment, determining a predicted respiratory feature value that matches the target respiratory feature value in the predicted respiratory waveform corresponding to the next cycle, so as to determine the timing of puncture in the next cycle, includes:

[0020] Obtain the fluctuation range of the respiratory characteristic value with the target respiratory characteristic value as the fluctuation benchmark;

[0021] The predicted respiratory feature values ​​that fall within the fluctuation range of the respiratory feature value in the predicted respiratory waveform corresponding to the next cycle are used as the predicted respiratory feature values ​​that match the target respiratory feature value, so as to determine the puncture timing in the next cycle.

[0022] In one embodiment, the method further includes: during the puncture process, if the predicted respiratory waveform corresponding to the next cycle at the current moment meets the preset respiratory abnormality conditions, sending a puncture termination command to the surgical robot; the puncture termination command is used to instruct the surgical robot to stop the puncture and perform a needle withdrawal operation.

[0023] In one embodiment, the step of sending a termination command to the surgical robot when the predicted respiratory waveform corresponding to the next cycle at the current moment meets a preset respiratory abnormality condition includes:

[0024] Determine the threshold range of the predicted feature values ​​representing the predicted respiratory waveform;

[0025] If the target respiratory feature value is not within the predicted feature value threshold range, the puncture termination command is sent to the surgical robot.

[0026] In one embodiment, the step of sending a termination command to the surgical robot when the predicted respiratory waveform corresponding to the next cycle at the current moment meets a preset respiratory abnormality condition includes:

[0027] Determine the actual feature value threshold range of the actual respiratory waveform representation corresponding to each of the previous several cycles of the next cycle, and determine the predicted feature value threshold range of the predicted respiratory waveform representation corresponding to the next cycle;

[0028] The average value of the actual feature value threshold interval corresponding to each of the aforementioned periods is determined to obtain the average actual feature value threshold interval;

[0029] When the difference between the predicted feature value threshold range and the average actual feature value threshold range meets a preset difference condition, the puncture termination command is sent to the surgical robot.

[0030] In one embodiment, the step of sending a termination command to the surgical robot when the predicted respiratory waveform corresponding to the next cycle at the current moment meets a preset respiratory abnormality condition includes:

[0031] When the difference between the predicted cycle duration and the actual cycle duration corresponding to the predicted respiratory waveform in the next cycle meets a preset difference condition, the puncture termination command is sent to the surgical robot.

[0032] Secondly, this application also provides a device for predicting the timing of puncture. The device includes:

[0033] The waveform acquisition module is used to acquire the actual respiratory waveform of the punctured object in the preceding several cycles of the next cycle, for the upcoming next cycle.

[0034] The input module is used to input the actual respiratory waveforms corresponding to the previous several cycles into the pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform corresponding to the next cycle.

[0035] The feature value acquisition module is used to acquire the target respiratory feature value of the punctured object; the target respiratory feature value is the respiratory feature value of the punctured object when acquiring the medical scan image used for needle tract planning of the puncture needle.

[0036] The determination module is used to determine a predicted respiratory feature value that matches the target respiratory feature value in the predicted respiratory waveform corresponding to the next cycle, so as to determine the puncture timing in the next cycle.

[0037] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to perform the following steps:

[0038] For the upcoming next cycle, obtain the actual respiratory waveform of the punctured subject in the previous several cycles of the next cycle;

[0039] The actual respiratory waveforms corresponding to the previous several cycles are input into the pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform for the next cycle.

[0040] Obtain the target respiratory feature value of the punctured object; the target respiratory feature value is the respiratory feature value of the punctured object when acquiring the medical scan image used for needle path planning of the puncture needle;

[0041] In the predicted respiratory waveform corresponding to the next cycle, a predicted respiratory feature value that matches the target respiratory feature value is determined in order to determine the timing of puncture in the next cycle.

[0042] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, performs the following steps:

[0043] For the upcoming next cycle, obtain the actual respiratory waveform of the punctured subject in the previous several cycles of the next cycle;

[0044] The actual respiratory waveforms corresponding to the previous several cycles are input into the pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform for the next cycle.

[0045] Obtain the target respiratory feature value of the punctured object; the target respiratory feature value is the respiratory feature value of the punctured object when acquiring the medical scan image used for needle path planning of the puncture needle;

[0046] In the predicted respiratory waveform corresponding to the next cycle, a predicted respiratory feature value that matches the target respiratory feature value is determined in order to determine the timing of puncture in the next cycle.

[0047] Fifthly, this application also provides a surgical execution device. The surgical execution device includes a medical image acquisition device, a respiratory motion acquisition device, a surgical robot, and a surgical planning and execution workstation; the medical image acquisition device is used to acquire medical scan images of the punctured object and send them to the surgical planning and execution workstation.

[0048] The surgical planning execution workstation is used to plan the puncture needle path based on the medical scan images;

[0049] The respiratory motion acquisition device is used to monitor the respiratory characteristic values ​​of the punctured object when the medical image acquisition device acquires medical scan images of the punctured object.

[0050] In preparation for the next cycle, the respiratory motion acquisition device is also used to acquire the actual respiratory waveform of the punctured object in the previous several cycles of the next cycle and send it to the surgical planning and execution workstation.

[0051] The surgical planning execution workstation is used to input the actual respiratory waveforms corresponding to the previous several cycles into a pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform for the next cycle.

[0052] The surgical planning and execution workstation is also used to take the respiratory feature value of the punctured object as the target respiratory feature value when acquiring the medical scan image of the punctured object, and determine the predicted respiratory feature value that matches the target respiratory feature value in the predicted respiratory waveform corresponding to the next cycle, so as to determine the puncture timing in the next cycle.

[0053] The surgical planning and execution workstation is also used to send a puncture command to the surgical robot when the puncture opportunity arrives;

[0054] The surgical robot is used to control the robotic arm to perform the puncture operation according to the puncture command.

[0055] In one embodiment, the surgical planning execution workstation is also used to perform the steps of implementing the above-described method.

[0056] The aforementioned method, device, computer equipment, storage medium, and surgical execution equipment for predicting puncture timing, since the most intuitive feature for predicting respiration is the respiratory waveform, acquire the actual respiratory waveform of the subject in the preceding several cycles for the upcoming cycle. Then, inputting the actual respiratory waveforms of the preceding cycles into a pre-trained respiratory waveform prediction model can more accurately obtain the predicted respiratory waveform for the next cycle. Next, acquire the target respiratory feature value of the subject; where the target respiratory feature value is the respiratory feature value of the subject when acquiring the medical scan image used for needle tract planning. Finally, compare the target respiratory feature value with the predicted respiratory feature value in the predicted respiratory waveform for the next cycle. By matching eigenvalues, a predicted respiratory characteristic value that matches the target respiratory characteristic value can be determined to determine the puncture timing in the next cycle. Since the puncture subject is in a stable respiratory state when acquiring the medical scan images used for needle path planning, meaning the target respiratory characteristic value is the respiratory characteristic value of the puncture subject in a stable respiratory state, the puncture timing predicted by the predicted respiratory characteristic value that matches the target respiratory characteristic value is closer to the stable respiratory state of the puncture subject. Therefore, when puncturing the subject according to this puncture timing, the actual target point reached by the interventional device can be closer to the target target point. Thus, predicting the puncture timing using the above method can effectively improve the accuracy of puncture timing prediction. Attached Figure Description

[0057] Figure 1 This is a flowchart illustrating a method for predicting the timing of puncture in one embodiment;

[0058] Figure 2 This is a schematic diagram of a respiratory motion waveform in one embodiment;

[0059] Figure 3 This is a schematic diagram of the structure of a BP neural network model in one embodiment;

[0060] Figure 4 This is a schematic diagram of the training process of a BP neural network model in one embodiment;

[0061] Figure 5 This is a flowchart illustrating a method for predicting the timing of puncture in another embodiment;

[0062] Figure 6 This is a schematic diagram of the structure of a surgical execution device in one embodiment;

[0063] Figure 7 This is a flowchart illustrating another method for predicting the timing of puncture in one embodiment;

[0064] Figure 8This is a structural block diagram of a device for predicting puncture timing in one embodiment;

[0065] Figure 9 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. Detailed Implementation

[0066] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0067] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0068] In one embodiment, such as Figure 1 As shown, a method for predicting the timing of puncture is provided. This embodiment illustrates the application of this method to a computer device, but it is understood that the method can also be applied to a server. The computer device can be, but is not limited to, various personal computers, laptops, smartphones, tablets, etc.; the server can be a standalone server or a server cluster composed of multiple servers. In this embodiment, the method includes the following steps:

[0069] Step S110: For the upcoming next cycle, obtain the actual respiratory waveform of the punctured object in the previous several cycles of the next cycle.

[0070] The upcoming next cycle is defined as the duration of one respiratory cycle following the current moment. For example, if the current moment is 4.1 seconds and the respiratory cycle duration is 1 second, then the upcoming next cycle will be 4.1 seconds to 5.1 seconds. Therefore, the first three cycles for the punctured subject in the next cycle could be 3.1 seconds to 4.1 seconds, 2.1 seconds to 3.1 seconds, and 1.1 seconds to 2.1 seconds. It's important to understand that the number of preceding cycles is not limited to three; it can be any number.

[0071] In practice, the most intuitive feature for predicting respiration is the respiratory waveform. During the operation, the respiratory motion acquisition device can collect the respiratory waveform of the punctured object in real time and send the respiratory waveform to the computer device. In this way, for the next cycle, the computer device can obtain the respiratory waveform of the punctured object in the previous several cycles of the next cycle, as the actual respiratory waveform of the previous several cycles.

[0072] Step S120: Input the actual respiratory waveforms corresponding to the previous several cycles into the pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform for the next cycle.

[0073] In practice, the computer device can input the actual respiratory waveforms corresponding to the previous few cycles into a pre-trained respiratory waveform prediction model, thereby predicting the respiratory waveform corresponding to the next cycle, which will then be used as the predicted respiratory waveform for the next cycle.

[0074] Understandably, the duration of the respiratory cycle varies from person to person and is not fixed. For example, if the current time is 4.5s, the first three cycles of the next cycle for the person being punctured could be 0s-1.6s, 1.6s-3.1s, and 3.1s-4.5s. The predicted cycle duration (duration of the respiratory cycle) for the predicted respiratory waveform in the next cycle, as predicted by the model, is 1.5s, meaning the model predicts the next cycle to be 4.5s-6s.

[0075] Step S130: Obtain the target respiratory feature value of the punctured object.

[0076] Among them, the target respiratory feature value is the respiratory feature value of the punctured object when acquiring the medical scan image used for needle tract planning.

[0077] Among these, respiratory characteristic values ​​are those that can represent respiration, that is, values ​​that can accurately reflect the respiratory waveform. The respiratory waveform corresponding to the respiratory characteristic value is a sine wave. Respiratory characteristic values ​​may include, but are not limited to, respiratory pressure values, respiratory flow rate values, etc.

[0078] Among them, medical scan images may include, but are not limited to, CT (Computed Tomography) images, MR (Magnetic Resonance) images, ultrasound images, and other medical images.

[0079] In practical applications, medical scan images can also be called preoperative images.

[0080] In practice, before the surgery, the computer equipment can acquire the respiratory characteristic values ​​of the patient during the medical scan image used for needle tract planning, which serves as the target respiratory characteristic value for the patient. Specifically, a respiratory motion acquisition device can acquire the respiratory characteristic values ​​of the patient during the preoperative image scan. During the preoperative image scan, the patient is asked to hold their breath, meaning they are in a stable breathing state. The respiratory characteristic values ​​measured by the respiratory motion acquisition device are recorded as the target respiratory characteristic value. This target respiratory characteristic value is not a fixed value for different patients or different scanning devices, but it is essential to ensure that the target respiratory characteristic value during the preoperative image scan matches the respiratory characteristic value during the puncture.

[0081] In practical applications, the volume of air inhaled by the subject during inhalation and breath-holding in a normal state is at its maximum (i.e., the respiratory pressure is at its maximum, which is the maximum pressure exerted on the chest pressure strap when data is collected). The respiratory flow rate is zero when the subject exhales all air and inhales completely. The target respiratory characteristic value used to predict the timing of puncture, obtained through the puncture timing prediction method, can theoretically be any value under stable respiratory conditions.

[0082] After obtaining medical scan images through preoperative imaging and computer equipment, users can plan their surgery on these images and use the plan as a reference during the procedure. This surgical planning can include designing the needle path for procedures such as biopsy and implantation. Once the planning is complete, the surgical procedure can begin.

[0083] Step S140: In the predicted respiratory waveform corresponding to the next cycle, determine the predicted respiratory feature value that matches the target respiratory feature value, so as to determine the puncture timing in the next cycle.

[0084] In practice, during the surgery, the computer device can match the target respiratory feature value with the predicted respiratory feature value in the predicted respiratory waveform of the next cycle. This allows the computer to determine the predicted respiratory feature value that matches the target respiratory feature value in the predicted respiratory waveform of the next cycle. Then, the computer can determine the puncture timing in the next cycle based on the time point corresponding to the matched predicted respiratory feature value. When the current moment is the puncture timing, the computer sends a puncture command to the surgical robot to start the puncture.

[0085] In the aforementioned method for predicting the timing of puncture, since the most intuitive feature for predicting respiration is the respiratory waveform, the actual respiratory waveform of the punctured subject in the preceding several cycles is obtained for the upcoming cycle. Then, the actual respiratory waveforms of the preceding cycles are input into a pre-trained respiratory waveform prediction model to more accurately obtain the predicted respiratory waveform for the next cycle. Next, the target respiratory feature value of the punctured subject is obtained; this target respiratory feature value is the respiratory feature value of the punctured subject when the medical scan image used for needle tract planning is acquired; specifically, the punctured subject is in a stable respiratory state when the medical scan image used for needle tract planning is acquired. Finally, the target respiratory feature value is compared with the predicted respiratory waveform for the next cycle. By matching the predicted respiratory feature value in the model with the target respiratory feature value, the predicted respiratory feature value that matches the target respiratory feature value can be determined to determine the puncture timing in the next cycle. In this way, since the puncture subject is in a stable respiratory state when the medical scan image used for needle path planning is acquired, that is, the target respiratory feature value is the respiratory feature value of the puncture subject in a stable respiratory state, the puncture timing predicted by the predicted respiratory feature value that matches the target respiratory feature value is closer to the respiratory stable state of the puncture subject. Therefore, when the puncture subject is punctured according to the puncture timing, the actual target point reached by the interventional device can be closer to the target target point. Therefore, the puncture timing can be predicted by the above method, which can effectively improve the prediction accuracy of puncture timing.

[0086] In one embodiment, the actual respiratory waveforms corresponding to the previous several cycles are input into a pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform for the next cycle, including: discretizing the continuous actual respiratory waveforms corresponding to the previous several cycles to obtain discrete actual respiratory waveforms corresponding to the previous several cycles; and inputting the discrete actual respiratory waveforms corresponding to the previous several cycles into a pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform for the next cycle.

[0087] The continuous actual respiratory waveform can be represented by a respiratory motion state waveform diagram, where the horizontal axis represents time and the vertical axis represents respiratory characteristic values. For ease of understanding by those skilled in the art, Figure 2 A schematic diagram of the respiratory motion state waveform is provided when the respiratory characteristic value is the respiratory pressure value. Among them, Figure 2 The respiratory pressure value corresponding to the dashed line in the image is the respiratory pressure value of the punctured object when the medical scan image is acquired, which is the target respiratory pressure value.

[0088] In practice, when the computer device inputs the actual respiratory waveforms corresponding to the previous several cycles into a pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform for the next cycle, the computer device can discretize the continuous actual respiratory waveforms corresponding to the previous several cycles to obtain discrete actual respiratory waveforms corresponding to the previous several cycles. Specifically, for a certain cycle within the previous several cycles, after discretizing the corresponding continuous actual respiratory waveform, each discrete time point and the corresponding actual respiratory feature value on the continuous actual respiratory waveform at each discrete time point can be obtained. Therefore, a set of corresponding discrete time points and actual respiratory feature values ​​for that cycle will serve as the discrete actual respiratory waveform for that cycle. Similarly, by discretizing the continuous actual respiratory waveform corresponding to each cycle within the previous several cycles using the same method, discrete actual respiratory waveforms corresponding to each cycle within the previous several cycles can be obtained, thus yielding the discrete actual respiratory waveforms corresponding to the previous several cycles.

[0089] For example, assuming the current time is 2.5s, the duration of the respiratory cycle is 1s, and a certain cycle is 1.5s to 2.5s, and one cycle is discretized into 11 discrete time points (the number of discrete time points can also be other numbers, which is not specifically limited here), then after discretizing the continuous actual respiratory waveform corresponding to this cycle, we can obtain the discrete time points 1.5s, 1.6s, 1.7s, 1.8s, 1.9s, 2.0s, 2.1s, 2.2s, 2.3s, 2.4s, and 2.5s, as well as the actual respiratory characteristic values ​​corresponding to these discrete time points on the continuous actual respiratory waveform corresponding to the 1.5s to 2.5s cycle. Then, the above discrete time points and the corresponding actual respiratory characteristic values ​​will be used as the discrete actual respiratory waveform for the 1.5s to 2.5s cycle.

[0090] Then, the computer device inputs the discrete actual respiratory waveforms corresponding to the previous few cycles into the pre-trained respiratory waveform prediction model, and can obtain the predicted respiratory waveform for the next cycle.

[0091] The technical solution of this embodiment discretizes the continuous actual respiratory waveforms corresponding to the previous several cycles to obtain discrete actual respiratory waveforms corresponding to the previous several cycles. The discrete actual respiratory waveforms corresponding to the previous several cycles are then input into a pre-trained respiratory waveform prediction model, which can quickly and accurately predict the predicted respiratory waveform corresponding to the next cycle based on historical respiratory waveform data.

[0092] In one embodiment, the discrete actual respiratory waveforms corresponding to the previous several cycles are input into a pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform for the next cycle, including: inputting the discrete actual respiratory waveforms corresponding to the previous several cycles into a pre-trained respiratory waveform prediction model to obtain the discrete predicted respiratory waveform for the next cycle; and performing interpolation processing on the discrete predicted respiratory waveform for the next cycle to obtain the predicted respiratory waveform for the next cycle.

[0093] In practice, the computer device inputs the discrete actual respiratory waveforms corresponding to the previous several cycles into a pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform for the next cycle. Then, the computer device can perform interpolation processing on the discrete predicted respiratory waveform for the next cycle to obtain a complete continuous respiratory waveform, which is used as the predicted respiratory waveform for the next cycle.

[0094] The technical solution of this embodiment is to input the discrete actual respiratory waveforms corresponding to the previous several cycles into a pre-trained respiratory waveform prediction model to obtain the discrete predicted respiratory waveforms corresponding to the next cycle; and to perform interpolation processing on the discrete predicted respiratory waveforms corresponding to the next cycle to obtain the predicted respiratory waveforms corresponding to the next cycle; in this way, the complete continuous respiratory waveforms corresponding to the next cycle can be obtained, so that the puncture timing can be determined more accurately in the complete continuous respiratory waveforms corresponding to the next cycle.

[0095] In one embodiment, in the predicted respiratory waveform corresponding to the next cycle, a predicted respiratory feature value matching the target respiratory feature value is determined to determine the puncture timing in the next cycle, including: obtaining the fluctuation range of the respiratory feature value with the target respiratory feature value as the fluctuation benchmark; and taking the predicted respiratory feature value falling within the fluctuation range of the respiratory feature value in the predicted respiratory waveform corresponding to the next cycle as the predicted respiratory feature value matching the target respiratory feature value to determine the puncture timing in the next cycle.

[0096] In practice, the computer device determines the predicted respiratory feature value that matches the target respiratory feature value in the predicted respiratory waveform corresponding to the next cycle, so as to determine the puncture timing in the next cycle. During this process, the computer device can obtain the fluctuation range of the respiratory feature value with the target respiratory feature value as the fluctuation benchmark. The predicted respiratory feature value that falls within the fluctuation range of the respiratory feature value in the predicted respiratory waveform corresponding to the next cycle is taken as the predicted respiratory feature value that matches the target respiratory feature value. The puncture timing is determined in the next cycle by using the matched predicted respiratory feature value. When the current moment is the puncture timing, the computer device sends a puncture command to the surgical robot to start the puncture.

[0097] The technical solution of this embodiment obtains the fluctuation range of respiratory feature values ​​with the target respiratory feature value as the fluctuation benchmark; then, the predicted respiratory feature values ​​that fall within the fluctuation range of respiratory feature values ​​in the predicted respiratory waveform corresponding to the next cycle are taken as the predicted respiratory feature values ​​that match the target respiratory feature value. In this way, the predicted respiratory feature value that is consistent with the target respiratory feature value corresponding to the puncture object in a stable respiratory state can be determined in the predicted respiratory waveform corresponding to the next cycle. Thus, the puncture timing can be determined in the next cycle based on the time point corresponding to the predicted respiratory feature value. This makes the respiratory state of the puncture object closer to a stable respiratory state when the puncture object is punctured according to the predicted puncture timing, thereby reducing the impact of organ deformation caused by breathing during puncture and making the puncture more accurate.

[0098] In one embodiment, the method further includes: during the puncture process, if the predicted respiratory waveform corresponding to the next cycle is detected to meet the preset respiratory abnormality conditions, sending a puncture termination command to the surgical robot; the puncture termination command is used to instruct the surgical robot to stop the puncture and perform a needle withdrawal operation.

[0099] In practice, doctors typically take certain measures to immobilize the patient during the puncture to prevent displacement that could affect the accuracy and safety of the procedure. However, unexpected situations such as severe coughing or pain can occur. Therefore, during the puncture, the respiratory motion acquisition device still collects respiratory waveforms and sends them to a computer for dynamic real-time respiratory waveform prediction. The computer predicts the respiratory waveform for the next cycle at each moment. When drastic respiratory changes cause a significant discrepancy between the actual respiratory characteristics and the predicted respiratory waveform, the computer determines that the predicted respiratory waveform for the next cycle meets preset abnormal respiratory conditions, indicating a puncture abnormality. In this case, the computer sends a puncture termination command to the surgical robot, instructing it to stop the puncture and withdraw the needle.

[0100] The technical solution of this embodiment improves safety during the puncture process by sending a puncture termination command to the surgical robot when the predicted respiratory waveform for the next cycle is detected to meet the preset abnormal respiratory conditions. The puncture termination command instructs the surgical robot to stop the puncture and perform needle withdrawal.

[0101] In one embodiment, when the predicted respiratory waveform corresponding to the next cycle at the current moment meets the preset respiratory abnormality conditions, a puncture termination command is sent to the surgical robot, including: determining the predicted feature value threshold range of the predicted respiratory waveform; and sending a puncture termination command to the surgical robot when the target respiratory feature value is not within the predicted feature value threshold range.

[0102] In practice, when the computer device detects that the predicted respiratory waveform corresponding to the next cycle at the current moment meets the preset respiratory abnormality conditions, it sends a puncture termination command to the surgical robot. During the puncture process, the computer device can determine the predicted feature value threshold range represented by the predicted respiratory waveform based on the maximum and minimum predicted respiratory feature values ​​in the predicted respiratory waveform corresponding to the next cycle. If the target respiratory feature value is not within the predicted feature value threshold range, it determines that there is a strong discrepancy between the actual respiratory feature value and the predicted respiratory waveform, that is, it detects that the predicted respiratory waveform corresponding to the next cycle meets the preset respiratory abnormality conditions, and then sends a puncture termination command to the surgical robot.

[0103] The technical solution of this embodiment determines the threshold range of predicted feature values ​​representing the predicted respiratory waveform; when the target respiratory feature value is not within the threshold range of the predicted feature value, a puncture termination command is sent to the surgical robot; thus, by comparing with the target respiratory feature value, it is possible to accurately determine whether the punctured object has respiratory abnormalities, thereby sending a puncture termination command in abnormal situations to effectively improve the safety of the puncture process.

[0104] In one embodiment, when the predicted respiratory waveform corresponding to the next cycle at the current moment meets a preset respiratory abnormality condition, a puncture termination command is sent to the surgical robot. This includes: determining the actual feature value threshold range of the actual respiratory waveform representation in each of the previous several cycles of the next cycle, and determining the predicted feature value threshold range of the predicted respiratory waveform representation in the next cycle; determining the average value of the actual feature value threshold ranges corresponding to each cycle to obtain the average actual feature value threshold range; and when the difference between the predicted feature value threshold range and the average actual feature value threshold range meets a preset difference condition, a puncture termination command is sent to the surgical robot.

[0105] In specific implementation, when the computer device detects that the predicted respiratory waveform corresponding to the next cycle at the current moment meets the preset respiratory abnormality conditions, and sends a puncture termination command to the surgical robot, the computer device can determine the actual feature value threshold range of the actual respiratory waveform corresponding to each of the previous several cycles during the puncture process, and determine the predicted feature value threshold range of the predicted respiratory waveform corresponding to the next cycle; then, based on the average value of the actual feature value threshold range corresponding to each of the previous several cycles, determine the average actual feature value threshold range; then, compare the predicted feature value threshold range with the average actual feature value threshold range, and when the difference between the predicted feature value threshold range and the average actual feature value threshold range meets the preset difference condition, it is determined that there is a strong discrepancy between the actual respiratory feature value and the predicted respiratory waveform, that is, the predicted respiratory waveform corresponding to the next cycle meets the preset respiratory abnormality conditions, and then a puncture termination command is sent to the surgical robot.

[0106] The technical solution of this embodiment determines the actual feature value threshold range of the actual respiratory waveform representation in each of the previous several cycles of the next cycle, and determines the predicted feature value threshold range of the predicted respiratory waveform representation in the next cycle; determines the average value of the actual feature value threshold ranges corresponding to each cycle, and obtains the average actual feature value threshold range; when the difference between the predicted feature value threshold range and the average actual feature value threshold range meets a preset difference condition, a puncture termination command is sent to the surgical robot; thus, when the predicted respiratory waveform obtained based on the actual respiratory waveforms corresponding to the previous several cycles has a large difference from the actual respiratory waveforms corresponding to the previous several cycles, it is determined that the puncture subject has a respiratory abnormality, thereby sending a puncture termination command to effectively improve the safety of the puncture process.

[0107] In one embodiment, when the predicted respiratory waveform corresponding to the next cycle at the current moment meets a preset respiratory abnormality condition, a puncture termination command is sent to the surgical robot, including: when the difference between the predicted cycle duration and the actual cycle duration corresponding to the predicted respiratory waveform in the next cycle meets a preset difference condition, a puncture termination command is sent to the surgical robot.

[0108] In practice, when the computer device detects that the predicted respiratory waveform corresponding to the next cycle at the current moment meets the preset respiratory abnormality conditions, it sends a puncture termination command to the surgical robot. During this process, the computer device can determine the predicted cycle length corresponding to the predicted respiratory waveform of the next cycle. When the difference between the predicted cycle length and the actual cycle length of the next cycle meets the preset difference conditions, it is determined that there is a strong discrepancy between the actual respiratory characteristic value and the predicted respiratory waveform. That is, when the predicted respiratory waveform corresponding to the next cycle meets the preset respiratory abnormality conditions, the computer device sends a puncture termination command to the surgical robot.

[0109] Specifically, if the difference between the predicted cycle duration and the actual cycle duration for the next cycle is greater than a preset cycle change threshold, the difference between the predicted cycle duration and the actual cycle duration is determined to meet the preset difference condition.

[0110] For example, if the current time is 4.5s, the duration of the puncture in the preceding cycles could be 0s-1.6s, 1.6s-3.1s, or 3.1s-4.5s. If the actual respiratory waveforms corresponding to these preceding cycles are input into a pre-trained respiratory waveform prediction model, the predicted duration of the next cycle is 4.5-6s. If the actual next cycle is 4.5-5.5s, then the predicted cycle duration is 1.5s, and the actual cycle duration is 1s. The difference between the predicted and actual cycle durations is 1.5s-1s = 0.5s. If the preset cycle change threshold is 0.3s, since the difference of 0.5s is greater than 0.3s, the difference between the predicted and actual cycle durations is determined to meet the preset difference condition. That is, if the predicted respiratory waveform for the next cycle meets the preset respiratory abnormality condition, a puncture termination command is sent to the surgical robot.

[0111] The technical solution of this embodiment sends a puncture termination command to the surgical robot when the difference between the predicted cycle length and the actual cycle length corresponding to the predicted respiratory waveform in the next cycle meets a preset difference condition. In this way, by comparing the predicted cycle length obtained based on historical respiratory waveform data with the actual cycle length, it is possible to accurately determine whether the punctured object has respiratory abnormalities, and send a puncture termination command in case of abnormalities to effectively improve the safety of the puncture process.

[0112] In one embodiment, the method further includes: using the period corresponding to the actual respiratory sample waveform as the label as the label period; training the respiratory waveform prediction model to be trained based on the actual respiratory sample waveforms corresponding to the first few periods of the label period and the actual respiratory sample waveforms corresponding to the label period, to obtain a pre-trained respiratory waveform prediction model.

[0113] Among them, the breathing waveform prediction model to be trained can be a BP neural network model.

[0114] In practice, before predicting the timing of the puncture, the computer device can acquire the actual respiratory waveforms of several puncture subjects, forming an actual respiratory waveform set. Then, the computer device can use the periods corresponding to the actual respiratory waveforms in the set as labels, and train the respiratory waveform prediction model based on the actual respiratory waveforms corresponding to the first few periods of the label period, thus obtaining a pre-trained respiratory waveform prediction model.

[0115] The structure diagram of a typical three-layer BP neural network model is as follows: Figure 3 As shown, X1, X2, X3, and X4 are input layer neurons, W1, W2, W3, and W4 are hidden layer neurons, and Y1 is the output.

[0116] When designing a BP neural network model, both the accuracy of the results and the prediction time requirements of the model should be considered. Therefore, the following aspects need to be taken into account:

[0117] 1. Selection and preprocessing of input features: The effectiveness of input feature values ​​determines the accuracy of the model to a certain extent. Although a large number of input features may improve the accuracy, they will inevitably reduce the prediction efficiency of the model. Therefore, it is necessary to find the most suitable effective input features. Sometimes these features are not obvious. In this case, it is necessary to preprocess the original data to obtain more effective features.

[0118] 2. During training, it is necessary to control the number of training iterations to avoid overfitting of the neural network model due to too many training iterations on the current dataset, or underfitting of the neural network model due to too few training iterations on the dataset.

[0119] 3. When tuning hyperparameters, use methods such as grid search to adjust hyperparameters such as learning rate, number of hidden layers, and activation function, and find the optimal hyperparameter configuration for the current input features while avoiding underfitting and overfitting.

[0120] 4. Finally, when evaluating the neural network design, it is necessary to consider the model's accuracy, the prediction time and computer resource consumption which are very important for real-time systems, as well as training costs and other factors. By taking all these factors into account, the optimal input features and the optimal hyperparameter configuration under those input features can be obtained.

[0121] The selection of input features and the adjustment of hyperparameters for the BP neural network algorithm not only ensure the algorithm's time efficiency but also save surgical time. In addition, the simplification of parameters can greatly reduce its consumption of computer resources.

[0122] The process involves training a pre-trained respiratory waveform prediction model based on the actual respiratory sample waveforms corresponding to the first few cycles of the label period and the actual respiratory sample waveforms corresponding to the label period. This includes: a computer inputting the actual respiratory sample waveforms corresponding to the first few cycles of the label period into the respiratory waveform prediction model for forward propagation to obtain the predicted respiratory sample waveforms corresponding to the label period; and updating the model parameters of the respiratory waveform prediction model for the label period through backpropagation based on the difference between the predicted respiratory sample waveforms and the corresponding actual respiratory sample waveforms to obtain the pre-trained respiratory waveform prediction model. When updating the model parameters, the weights of the connections between each neuron are updated.

[0123] For the ease of understanding of those skilled in the art, Figure 4 A schematic diagram of the training process for a BP neural network model is provided, such as... Figure 4 As shown, the actual respiratory sample waveform is a discrete actual respiratory sample waveform. The computer device can use the discrete actual respiratory sample waveforms corresponding to the first few periods of the label period, such as the discrete actual respiratory sample waveforms corresponding to the first three periods (including the one-to-one correspondence of discrete time points and actual respiratory sample feature values) as the input features of the BP neural network. The output of the BP neural network is the discrete predicted respiratory sample waveform corresponding to the label period. The discrete predicted respiratory sample waveform is interpolated to obtain the complete predicted respiratory waveform corresponding to the label period, which is used as the predicted respiratory sample waveform corresponding to the label period. Based on the difference between the predicted respiratory sample waveform corresponding to the label period and the corresponding actual respiratory sample waveform, backpropagation is used to determine the response error of the hidden layer and output layer of the BP neural network. The weights of each neuron of the BP neural network are updated through backpropagation. Finally, training is stopped when the training requirements are met, and the pre-trained respiratory waveform prediction model is obtained.

[0124] The input features are not limited to discrete actual respiratory sample waveforms corresponding to the first three cycles, but more input features will reduce the efficiency of the neural network. It is necessary to find a balance between the number of input features and the efficiency of the neural network.

[0125] In practical applications, during the process of determining the actual respiratory sample waveform as a label, the computer device can acquire discrete raw respiratory sample waveforms corresponding to several punctured sample objects collected by the respiratory motion acquisition device; then, the discrete raw respiratory sample waveforms are filtered to obtain the actual respiratory sample waveform set; then, the computer device can determine the actual respiratory sample waveform as a label from the actual respiratory sample waveform set.

[0126] In the process of filtering discrete raw respiratory sample waveforms to obtain the actual respiratory sample waveform set, the computer equipment can remove sample waveforms that meet the invalid data conditions, resulting in filtered discrete raw respiratory sample waveforms. For example, invalid sample waveform data caused by incorrect wearing of the respiratory motion acquisition device can be removed. Then, the computer equipment can preprocess the filtered discrete raw respiratory sample waveforms (e.g., data trimming) to obtain processed discrete raw respiratory sample waveforms. Then, the computer equipment can classify the processed discrete raw respiratory sample waveforms according to the sample waveform acquisition scenario, and obtain the actual respiratory sample waveform set based on the classified discrete raw respiratory sample waveforms, which can be used to train the BP neural network model.

[0127] The processed discrete raw respiratory sample waveforms can be classified into different scenarios, such as peaceful conditions and surgical conditions, making the distribution of training data more reasonable. When training the model, a portion of the dataset is selected for training. Too much peaceful data can lead to poor prediction performance for surgical conditions, and vice versa. Therefore, scenario classification is necessary to achieve a balance of input data.

[0128] During training, the actual respiratory sample waveform set can be reasonably divided into training, testing, and validation sets. After this division, the computer equipment can design a suitable backpropagation (BP) neural network model based on the actual situation. The number of input layer neurons, hidden layer neurons, the number of hidden layers, and the learning rate, among other hyperparameters, are roughly determined. Then, the model is trained using the training set, and the training results are tested using the testing set. The hyperparameters are adjusted based on the test results to obtain the optimal parameters for the best test results. Next, the model is validated using the validation set, and evaluation metrics such as accuracy and processing time are obtained to assess the model. Finally, the final neural network model is integrated into the application for puncture timing prediction.

[0129] The technical solution in this embodiment utilizes a backpropagation (BP) neural network to predict the optimal puncture time in real time. This transforms the traditional method of determining puncture timing from subjective judgment by the physician to active prediction by the surgical robot based on the actual respiratory waveforms collected by the respiratory motion acquisition device. This fulfills the prerequisite for automating the puncture process in interventional surgery robots. The application of the BP neural network learning algorithm can quickly, accurately, and robustly predict the optimal puncture timing based on the actual respiratory waveforms corresponding to the previous several cycles. By predicting the puncture timing in advance, not only can the accuracy and precision of the puncture be improved, but the foundation for automating the puncture procedure can also be laid.

[0130] In another embodiment, such as Figure 5 As shown, a method for predicting the timing of puncture is provided. Taking the application of this method to a computer device as an example, the method includes the following steps:

[0131] Step S502: For the upcoming next cycle, obtain the continuous actual respiratory waveform of the punctured object in the previous several cycles of the next cycle;

[0132] Step S504: Discretize the continuous actual respiratory waveforms corresponding to the first few cycles to obtain the discrete actual respiratory waveforms corresponding to the first few cycles.

[0133] Step S506: Input the discrete actual respiratory waveforms corresponding to the previous several cycles into the pre-trained respiratory waveform prediction model to obtain the discrete predicted respiratory waveforms corresponding to the next cycle.

[0134] Step S508: Interpolate the discrete predicted respiratory waveform corresponding to the next cycle to obtain the predicted respiratory waveform corresponding to the next cycle.

[0135] Step S510: Obtain the target respiratory characteristic value of the punctured object as the fluctuation range of the respiratory characteristic value based on the fluctuation benchmark.

[0136] Step S512: The predicted respiratory feature value that falls within the fluctuation range of the respiratory feature value in the predicted respiratory waveform corresponding to the next cycle is taken as the predicted respiratory feature value that matches the target respiratory feature value, so as to determine the puncture timing in the next cycle.

[0137] It should be noted that the specific limitations of the above steps can be found in the specific limitations of a method for predicting the timing of puncture described above.

[0138] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0139] Based on the same inventive concept, embodiments of this application also provide a surgical execution device for implementing the above-mentioned method for predicting the timing of puncture. For example... Figure 6 As shown, the surgical execution equipment includes medical image acquisition equipment, respiratory motion acquisition equipment, surgical robot, and surgical planning and execution workstation;

[0140] Medical image acquisition equipment is used to acquire medical scan images of the punctured object and send them to the surgical planning and execution workstation;

[0141] The surgical planning and execution workstation is used to plan the puncture needle path based on medical scan images. Specifically, the medical scan images are used by the user to plan the puncture needle path to determine the puncture path, target points, and other information as a reference during the surgery.

[0142] A respiratory motion acquisition device is used to monitor the respiratory motion state of a punctured object when a medical image acquisition device acquires a medical scan image of the punctured object, so as to obtain the target respiratory feature value of the punctured object.

[0143] In preparation for the upcoming next cycle, the respiratory motion acquisition device is also used to acquire the actual respiratory waveforms of the punctured subject in the previous several cycles of the next cycle and send them to the surgical planning and execution workstation.

[0144] The surgical planning and execution workstation is used to input the actual respiratory waveforms corresponding to the previous several cycles into the pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform for the next cycle.

[0145] The surgical planning and execution workstation is also used to take the respiratory feature value of the puncture subject as the target respiratory feature value when acquiring medical scan images of the puncture subject, and determine the predicted respiratory feature value that matches the target respiratory feature value in the predicted respiratory waveform corresponding to the next cycle, so as to determine the puncture timing in the next cycle.

[0146] The surgical planning and execution workstation is also used to send puncture instructions to the surgical robot when the puncture opportunity arrives;

[0147] The surgical robot is used to control the robotic arm to perform puncture operations according to puncture instructions, and to report the relevant status of the punctured object to the surgical planning and execution workstation during the puncture process.

[0148] The surgical planning execution workstation is also used to execute the steps of the aforementioned method for predicting the timing of puncture.

[0149] In addition, during the procedure, the respiratory motion acquisition device is also used to acquire the real-time phase and respiratory motion status of the punctured subject, so as to obtain the actual respiratory waveform of the punctured subject during the procedure.

[0150] In practical applications, the surgical planning and execution workstation can be a computer device equipped with surgical planning and execution software.

[0151] For ease of understanding in this field, Figure 7 A flowchart illustrating a method for predicting the timing of a puncture using the aforementioned surgical device is provided. The method employs respiratory pressure as a respiratory characteristic value and uses a CT scanner to obtain a preoperative image. In this embodiment, the method includes the following steps:

[0152] 1. The patient undergoes a preoperative CT scan wearing a respiratory motion acquisition device with a pressure sensor. During the scan, the respiratory motion acquisition device simultaneously records the pressure value at that time, which is used as the target respiratory pressure value. The target respiratory pressure value is then sent to the surgical planning and execution workstation.

[0153] 2. The user uses the preoperative diagram to plan the puncture needle path on the surgical planning and execution workstation. After the planning is completed, the surgical procedure begins.

[0154] 3. During surgery, the surgical planning execution workstation runs a pre-trained respiratory waveform prediction model. It predicts the respiratory waveform for the next cycle by acquiring the actual respiratory waveform in real time using respiratory motion acquisition equipment, and identifies the puncture opportunity based on the target respiratory pressure value from the preoperative scan. If the current time is opportune for puncture, a puncture command is sent to the robot to begin the puncture. In practical applications, the respiratory pressure value measured by the chest strap is set between 1 kPa and 5 kPa, without affecting the normal breathing of the patient.

[0155] 4. During the puncture, the respiratory motion acquisition device continues to collect respiratory waveforms and send them to the surgical planning and execution workstation for respiratory waveform prediction. If the surgical planning and execution workstation determines that an abnormality has occurred during the puncture, it will send a puncture termination command to the surgical robot to execute the needle withdrawal operation. This provides protection against unexpected situations that may alter the patient's condition, such as coughing during the procedure, thereby simplifying the entire surgical process, improving surgical precision, and reducing patient trauma and postoperative recurrence rates.

[0156] Based on the same inventive concept, this application also provides a device for predicting puncture timing to implement the above-mentioned method for predicting puncture timing. The solution provided by this device is similar to the solution described in the above-described method. Therefore, the specific limitations of one or more embodiments of the device for predicting puncture timing provided below can be found in the limitations of the method for predicting puncture timing described above, and will not be repeated here.

[0157] In one embodiment, such as Figure 8 As shown, a device for predicting the timing of puncture is provided, comprising: a waveform acquisition module 810, an input module 820, a feature value acquisition module 830, and a determination module 840, wherein:

[0158] The waveform acquisition module 810 is used to acquire the actual respiratory waveform of the punctured object in the preceding several cycles of the next cycle, for the upcoming next cycle.

[0159] The input module 820 is used to input the actual respiratory waveforms corresponding to the previous several cycles into the pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform corresponding to the next cycle.

[0160] The feature value acquisition module 830 is used to acquire the target respiratory feature value of the punctured object; the target respiratory feature value is the respiratory feature value of the punctured object when acquiring the medical scan image used for needle path planning of the puncture needle.

[0161] The determination module 840 is used to determine a predicted respiratory feature value that matches the target respiratory feature value in the predicted respiratory waveform corresponding to the next cycle, so as to determine the puncture timing in the next cycle.

[0162] In one embodiment, the input module 820 is specifically used to discretize the continuous actual respiratory waveforms corresponding to the previous several cycles to obtain discrete actual respiratory waveforms corresponding to the previous several cycles; and input the discrete actual respiratory waveforms corresponding to the previous several cycles into the pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveforms corresponding to the next cycle.

[0163] In one embodiment, the input module 820 is specifically used to obtain the discrete actual breathing waveform corresponding to the previous several cycles based on the actual breathing feature value corresponding to the continuous actual breathing waveform at discrete time points in the previous several cycles.

[0164] In one embodiment, the input module 820 is specifically used to input the discrete actual respiratory waveforms corresponding to the previous several cycles into the pre-trained respiratory waveform prediction model to obtain the discrete predicted respiratory waveform corresponding to the next cycle; and to perform interpolation processing on the discrete predicted respiratory waveform corresponding to the next cycle to obtain the predicted respiratory waveform corresponding to the next cycle.

[0165] In one embodiment, the determining module 840 is specifically used to obtain the fluctuation range of respiratory feature values ​​with the target respiratory feature value as the fluctuation benchmark; and to take the predicted respiratory feature values ​​that fall within the fluctuation range of the respiratory feature values ​​in the predicted respiratory waveform corresponding to the next cycle as the predicted respiratory feature values ​​that match the target respiratory feature value, so as to determine the puncture timing in the next cycle.

[0166] In one embodiment, the device further includes: an anomaly detection module, used to send a puncture termination command to the surgical robot when the predicted respiratory waveform corresponding to the next cycle at the current moment meets a preset respiratory anomaly condition during the puncture process; the puncture termination command is used to instruct the surgical robot to stop the puncture and perform a needle withdrawal operation.

[0167] In one embodiment, the anomaly detection module is specifically used to determine the predicted feature value threshold range of the predicted respiratory waveform representation; and when the target respiratory feature value is not within the predicted feature value threshold range, to send the end-puncture command to the surgical robot.

[0168] In one embodiment, the anomaly detection module is specifically used to determine the actual feature value threshold range of the actual respiratory waveform representation corresponding to each of the previous several cycles of the next cycle, and to determine the predicted feature value threshold range of the predicted respiratory waveform representation corresponding to the next cycle; to determine the average value of the actual feature value threshold range corresponding to each cycle, thereby obtaining the average actual feature value threshold range; and when the difference between the predicted feature value threshold range and the average actual feature value threshold range meets a preset difference condition, to send the puncture termination command to the surgical robot.

[0169] In one embodiment, the anomaly detection module is specifically used to send the end-puncture command to the surgical robot when the difference between the predicted cycle duration and the actual cycle duration corresponding to the predicted respiratory waveform of the next cycle meets a preset difference condition.

[0170] The modules in the aforementioned device for predicting the timing of puncture can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0171] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 9 As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a method for predicting the timing of puncture. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0172] Those skilled in the art will understand that Figure 9 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0173] In one embodiment, a computer device is further provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0174] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0175] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0176] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data shall comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0177] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0178] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0179] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A device for predicting the timing of puncture, characterized in that, The device includes: The waveform acquisition module is used to acquire the actual respiratory waveform of the punctured object in the preceding several cycles of the next cycle, for the upcoming next cycle. The input module is used to input the actual respiratory waveforms corresponding to the previous several cycles into the pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform corresponding to the next cycle. The feature value acquisition module is used to acquire the target respiratory feature value of the punctured object; the target respiratory feature value is the respiratory feature value of the punctured object when acquiring the medical scan image used for needle tract planning of the puncture needle. The determination module is used to determine a predicted respiratory feature value that matches the target respiratory feature value in the predicted respiratory waveform corresponding to the next cycle, so as to determine the puncture timing in the next cycle. An anomaly detection module is used to determine, during the puncture process, the actual feature value threshold range of the actual respiratory waveform representation corresponding to each of the previous several cycles of the next cycle at the current moment, and to determine the predicted feature value threshold range of the predicted respiratory waveform representation corresponding to the next cycle. The anomaly detection module is also used to determine the average value of the actual feature value threshold interval corresponding to each of the periods, so as to obtain the average actual feature value threshold interval. The anomaly detection module is further configured to send a puncture termination command to the surgical robot when the difference between the predicted feature value threshold range and the average actual feature value threshold range meets a preset difference condition; the puncture termination command is used to instruct the surgical robot to stop puncture and perform needle withdrawal operation.

2. The apparatus according to claim 1, characterized in that, The input module is specifically used to discretize the continuous actual breathing waveforms corresponding to the first few cycles to obtain the discrete actual breathing waveforms corresponding to the first few cycles. The discrete actual respiratory waveforms corresponding to the previous several cycles are input into the pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform for the next cycle.

3. The apparatus according to claim 2, characterized in that, The input module is specifically used to obtain the discrete actual breathing waveform corresponding to the previous several cycles based on the actual breathing characteristic values ​​corresponding to the continuous actual breathing waveform at discrete time points in the previous several cycles.

4. The apparatus according to claim 2, characterized in that, The input module is specifically used to input the discrete actual respiratory waveforms corresponding to the previous several cycles into the pre-trained respiratory waveform prediction model to obtain the discrete predicted respiratory waveform corresponding to the next cycle; and to perform interpolation processing on the discrete predicted respiratory waveform corresponding to the next cycle to obtain the predicted respiratory waveform corresponding to the next cycle.

5. The apparatus according to claim 1, characterized in that, The determining module is specifically used to obtain the fluctuation range of respiratory feature values ​​with the target respiratory feature value as the fluctuation benchmark; and to take the predicted respiratory feature values ​​that fall within the fluctuation range of the respiratory feature values ​​in the predicted respiratory waveform corresponding to the next cycle as the predicted respiratory feature values ​​that match the target respiratory feature value, so as to determine the puncture timing in the next cycle.

6. The apparatus according to claim 1, characterized in that, The anomaly detection module is also used to send the puncture termination command to the surgical robot when it detects that the predicted respiratory waveform corresponding to the next cycle at the current moment meets the preset respiratory anomaly conditions during the puncture process.

7. The apparatus according to claim 6, characterized in that, The anomaly detection module is specifically used to determine the predicted feature value threshold range of the predicted respiratory waveform representation; if the target respiratory feature value is not within the predicted feature value threshold range, the module sends the end puncture command to the surgical robot.

8. The apparatus according to claim 6, characterized in that, The anomaly detection module is specifically used to send the end-puncture command to the surgical robot when the difference between the predicted cycle duration and the actual cycle duration corresponding to the predicted respiratory waveform in the next cycle meets a preset difference condition.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it performs the steps performed by the apparatus of any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it performs the steps performed by the apparatus of any one of claims 1 to 8.

11. A surgical execution device, characterized in that, The equipment includes medical image acquisition equipment, respiratory motion acquisition equipment, surgical robot, and surgical planning and execution workstation; The medical image acquisition device is used to acquire medical scan images of the punctured object and send them to the surgical planning and execution workstation; The surgical planning execution workstation is used to plan the puncture needle path based on the medical scan images; The respiratory motion acquisition device is used to monitor the respiratory characteristic values ​​of the punctured object when the medical image acquisition device acquires medical scan images of the punctured object. In preparation for the next cycle, the respiratory motion acquisition device is also used to acquire the actual respiratory waveform of the punctured object in the previous several cycles of the next cycle and send it to the surgical planning and execution workstation. The surgical planning execution workstation is used to input the actual respiratory waveforms corresponding to the previous several cycles into a pre-trained respiratory waveform prediction model to obtain the predicted respiratory waveform for the next cycle. The surgical planning and execution workstation is also used to take the respiratory feature value of the punctured object as the target respiratory feature value when acquiring the medical scan image of the punctured object, and determine the predicted respiratory feature value that matches the target respiratory feature value in the predicted respiratory waveform corresponding to the next cycle, so as to determine the puncture timing in the next cycle. The surgical planning and execution workstation is also used to send a puncture command to the surgical robot when the puncture opportunity arrives; The surgical robot is used to control the robotic arm to perform the puncture operation according to the puncture command; The surgical planning execution workstation is also used to determine, during the puncture process, the actual feature value threshold range of the actual respiratory waveform representation corresponding to each cycle in the previous several cycles of the next cycle at the current moment, and to determine the predicted feature value threshold range of the predicted respiratory waveform representation corresponding to the next cycle. The surgical planning execution workstation is also used to determine the average value of the actual feature value threshold interval corresponding to each cycle, so as to obtain the average actual feature value threshold interval. The surgical planning execution workstation is also used to send a puncture termination command to the surgical robot when the difference between the predicted feature value threshold interval and the average actual feature value threshold interval meets a preset difference condition. The surgical robot is also used to stop the puncture and perform the needle withdrawal operation according to the puncture termination command.

12. The device according to claim 11, characterized in that, The surgical planning execution workstation is also used to perform the steps executed by the apparatus according to any one of claims 1 to 8.

Images