Construction method and application of cutting state classification model during human bone cutting by medical bone swing saw

Abstract

The invention discloses a construction method and application of a cutting state classification model during human bone cutting by a medical bone swing saw. The construction method comprises the following steps: S1, preparing an artificial cancellous bone as an experimental material; s2, classifying the cutting states; the three states are respectively a cutting preparation state, a pre-cutting state, a real-time cutting state, a critical state and a penetrating state; the acceleration signal characteristics of all the states are defined respectively, and the cutting process is determined according to vibration and frequency changes; s3, the sampling frequency of the acceleration signals is set to be 4000 Hz, every 64 ms is a sampling frame, and 15-dimensional features are obtained through feature extraction; the features include one-dimensional frequency domain features, five-dimensional wavelet features and nine-dimensional time domain features; s4, according to the fifteen-dimensional features obtained through feature extraction, the acceleration signals in the cutting process are manually divided into the five cutting states defined in the step S2, and the starting point and the ending point of each state are marked according to different feature changes; s5, establishing a classification model; a support vector machine SVM classification model is used, and a polynomial kernel function is adopted.

Description

Technical Field

[0001] The present application relates to the field of auxiliary bone surgery, and in particular to a method for constructing a cutting state classification model when a medical bone oscillating saw cuts human bones, and its application. Background Art

[0002] In orthopedic surgery, the medical bone oscillating saw is a common surgical tool used by surgeons, mainly for bone cutting. The saw blade of this tool produces a strong swing around the axis after the power is turned on. However, during the operation, the coolant and the shavings produced by the cutting will block the surgeon's field of vision, making it difficult to clearly observe the cutting area. Therefore, the surgeon mainly relies on listening to the sound produced when the tool contacts the bone tissue and the vibration felt by the hand to judge the cutting situation. For example, when the cutting goes smoothly, the frequency, intensity and rhythm changes of the vibration will be relatively small; when cutting into bone tissues of different densities, such as cortical bone and cancellous bone, the changes in the vibration rhythm will increase significantly.

[0003] This method relies on the subjective judgment of the doctor, but each doctor has different abilities to perceive vibrations. Some doctors can keenly capture subtle changes, while others may ignore this information. Therefore, for inexperienced surgeons, the accuracy of judging the cutting status by vibration is often low. For example, in total knee replacement surgery, if over-penetration occurs when using a bone oscillating saw, the carotid artery in the lower leg may be accidentally cut, posing a serious surgical risk. Summary of the invention

[0004] The purpose of the present invention is to overcome the deficiencies in the prior art, solve the inaccuracy problem caused by manual reliance on hand-sensing vibration for cutting state identification, and provide a method for constructing and applying a cutting state classification model when a medical bone oscillating saw cuts human bones.

[0005] The objective of the present invention is achieved through the following technical solutions:

[0006] A method for constructing a cutting state classification model when a medical bone oscillating saw cuts human bones, comprising: S1. Prepare artificial cancellous bone as experimental material; S2. Classify the cutting states into cutting preparation state, pre-cutting state, real-time cutting state, critical state and penetration state; define the acceleration signal characteristics of each state, and determine the cutting process based on the changes in vibration and frequency; S3. Set the sampling frequency of the acceleration signal to 4000 Hz, with each 64 ms as a sampling frame, and obtain 15-dimensional features through feature extraction, including one-dimensional frequency domain features, five-dimensional wavelet features, and nine-dimensional time domain features; S4. Based on the fifteen-dimensional features obtained by feature extraction, the acceleration signal during the cutting process is manually divided into the five cutting states defined in step S2, and the start and end points of each state are marked according to different feature changes; S5. Establish a classification model; use the support vector machine SVM classification model and adopt a polynomial kernel function. Furthermore, the experimental material of the artificial cancellous bone in step S1 is solid hard polyurethane foam plastic with a thickness of 4 mm. Furthermore, in step S2, the definitions of each state are as follows: In the cutting preparation state, the bone oscillating saw first accelerates to a predetermined speed and then approaches the experimental material to be cut; the acceleration signal in this state is caused by the reciprocating motion of the bone oscillating saw itself; The pre-cutting state marks the beginning of the cutting process, when the bone pendulum saw first contacts the experimental material and gradually cuts into it; as the experimental material deforms and the bone pendulum saw penetrates deeper, the interaction between the two causes vibrations; Real-time cutting state, during which the interaction between the bone pendulum saw and the experimental material reaches its peak; the acceleration signal in this state has a significant peak and oscillation mode, and the high-frequency component increases; The critical state occurs when the bone saw is about to cut out of the experimental material and the cutting process is nearing its end. The acceleration signal in this state shows a significant downward trend and is accompanied by fluctuations, indicating that the cutting activity has not been completely terminated. In the penetration state, the bone saw is completely separated from the experimental material and the cutting work is completed. Further, in step S3, One-dimensional frequency domain features: a sampling frame is transformed by FFT to obtain its spectrum diagram, and the frequency value corresponding to the maximum peak value of the sampling frame is found in the frequency range of 200Hz to 300Hz as its fundamental frequency value; the fundamental frequency value is the selected one-dimensional frequency domain feature; Five-dimensional wavelet features, using maximum overlap discrete wavelet transform MODWT to decompose the sampling frame and extract 5 relative energies as five-dimensional wavelet features; The nine-dimensional time domain features include nine statistical features introduced into the sampling frame: mean, variance, peak, effective value, peak factor, kurtosis, pulse factor, waveform factor, and margin factor. Furthermore, a Symlet wavelet of order 4 is used as a signal processing tool; the sampling frame is decomposed into 5 layers W1, W2, W3, W4, and W5 using MODWT, and the frequency range corresponding to each layer is 1000-2000Hz, 500-1000Hz, 250-500Hz, 125-250Hz, and 60-125Hz respectively; the signal energy corresponding to the decomposed W1, W2, W3, W4, and W5 layers is divided by the energy of the sampling frame signal to obtain five relative energies; these five relative energies are the five-dimensional wavelet features. Furthermore, the state calibration process of each complete cutting process in step S4 is as follows: Calibration in the cutting preparation state: In a complete cutting process, the amplitude of the spectrum graph of the cutting preparation state does not exceed 1, so the initial moment until the amplitude exceeds 1 is marked as the cutting preparation state; Pre-cutting state: the end of the cutting preparation marks the beginning of the pre-cutting state, and the end mark is that the fundamental frequency values ​​extracted from the three sampling frames are equal, marking the last sampling frame as the end of the pre-cutting state; Real-time cutting state: the middle part between the end of pre-cutting state and the beginning of critical state; Critical stage: the part of 3 sampling frames before the penetration state begins; Penetration state: After the bone pendulum saw penetrates, the saw blade will block the optical path of the laser sensor and transmit a high-level signal to the main control chip, and the high-level signal is used to mark the penetration state; The present invention also provides a device for identifying the cutting state when a medical bone swing saw cuts human bones, comprising a main control chip, a storage module, an accelerometer, a communication module, a laser sensor, a display module and a power module; The main control chip is respectively connected to the storage module, the accelerometer, the communication module, the laser sensor, the display module and the power module; The communication module uses Bluetooth transmission, the display module uses a serial port screen, the laser sensor is used to calibrate the penetration state, and the detection frequency is 100 Hz; the cutting direction of the bone pendulum saw is to cut downward perpendicular to the bone plate, and the collected acceleration signal is the vibration direction of the bone pendulum saw. The present invention also provides an application of a cutting state classification model when a medical bone oscillating saw cuts human bones, based on the method for constructing a cutting state classification model when a medical bone oscillating saw cuts human bones as described in any one of claims 1-6 and the cutting state recognition device of claim 7, characterized in that the accelerometer of the cutting state recognition device collects signals and calculates the fundamental frequency amplitude; features are extracted from the collected signals and input into a support vector machine (SVM) classification model for classification; and the classification results are displayed in real time to assist doctors in judging the cutting state. The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps of a method for constructing a cutting state classification model when the medical bone oscillating saw cuts human bones are implemented. The present invention also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the method for constructing a cutting state classification model when the medical bone oscillating saw cuts human bones are implemented.

[0007] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0008] 1. Improve the accuracy of cutting state recognition; the present invention is based on the feature extraction of acceleration signals and combines the SVM classification model for recognition. Compared with the traditional method that relies on the subjective perception of doctors, the cutting state is quantified through the objective characteristics of vibration and frequency, which reduces human errors. It is particularly suitable for inexperienced doctors and improves the accuracy of cutting state judgment.

[0009] 2. Reduce surgical risks: By accurately identifying the cutting state, especially the critical and penetrating states, the risk of over-cutting is effectively avoided, and the possibility of accidental injury to surrounding important blood vessels or nerves is reduced. This is particularly important in total knee replacement surgery, avoiding accidental injury to important tissues such as the calf carotid artery, thereby improving the safety of the operation.

[0010] 3. Improve surgical efficiency; automatic identification of cutting status can provide real-time feedback on the cutting progress, reducing the time doctors spend on observation or adjustment during the cutting process, helping to improve overall surgical efficiency and shorten the duration of surgery.

[0011] 4. Strong adaptability, able to cope with complex cutting processes: The present invention divides the cutting process into five states (cutting preparation, pre-cutting, real-time cutting, critical, and penetration), and defines the characteristic changes of each state. This segmentation method can comprehensively cover the different states in the cutting process, and helps to cope with the impact of changes in bone tissue density and hardness on vibration and acceleration, thereby achieving accurate identification of complex cutting processes.

[0012] 5. Facilitate data analysis and improvement; all cutting data and feature information are stored through the storage module (SD card), so that the acceleration signal characteristics, classification model performance and cutting conditions in each state can be analyzed after the operation, providing data support for future model optimization and equipment improvement, and further improving the reliability and applicability of recognition.

[0013] 6. Applicable to various types of bone tissues; through the combined extraction of frequency domain, wavelet and time domain features, the invention can meet the cutting needs of bone tissues of different densities, such as cortical bone and cancellous bone, and is suitable for various types of bone cutting, which is conducive to promotion and application in various orthopedic surgeries.

[0014] 7. Real-time display of cutting status to improve the doctor's operating experience; during the operation, the recognition device can display the cutting status to the doctor in real time, so that the doctor can intuitively grasp the progress of the operation, reduce dependence on hand vibration or sound judgment, and improve the surgical operation experience.

[0015] In summary, this invention combines machine learning models with sensor technology to construct an automated, precise, and real-time feedback bone swing saw cutting state recognition device, realizing precise surgical assistance functions, and has significant application value and promotion prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a structural schematic diagram of the present invention.

[0017] Figure 2 It is the state diagrams and their corresponding amplitude-frequency curves after the cutting states are classified.

[0018] Figure 3a is the sampling frame diagram; Figure 3b It is the signal diagram after the sampling frame is decomposed by MODWT.

[0019] Figure 4 This is a diagram of the confusion matrix of the test set.

[0020] Figure 5 It is a flow chart of the use of cutting state classification model. DETAILED DESCRIPTION

[0021] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0022] like Figure 1 As shown, the cutting state identification device for a medical bone swing saw when cutting human bones provided in this embodiment includes a main control chip, a storage module, an accelerometer, a communication module, a laser sensor, a display module and a power module; the main control chip is respectively connected to the storage module, the accelerometer, the communication module, the laser sensor, the display module and the power module; the power module is used to supply power to each module;

[0023] The main control chip used in this embodiment is STM32F407, the storage module is an SD card, which is used to store the acceleration signal during the implementation process, the accelerometer (ADXL375, ADI, USA) has a maximum range of 200g, and the communication module adopts Bluetooth transmission, the specific model is: JDY-34 (China, Yourcee); the laser sensor adopts LSZ-S30N1 (China, GTRIC); used to calibrate the penetration state, the detection frequency is 100Hz;

[0024] The display module uses a serial port screen (TJC3224T124-011N, China, TAOJINGCHI);

[0025] The cutting direction of the bone oscillating saw is vertical and downward to the bone plate, and the collected acceleration signal is the vibration direction of the bone oscillating saw.

[0026] Preferably, this embodiment provides a method for constructing a cutting state classification model when a medical bone oscillating saw cuts human bones, comprising:

[0027] S1. Prepare artificial cancellous bone as experimental material; the experimental material used in this embodiment is artificial cancellous bone (40PCF, SAWBONES, USA), which is made of solid hard polyurethane foam plastic with a thickness of 4 mm.

[0028] S2. Classify the cutting states into cutting preparation state, pre-cutting state, real-time cutting state, critical state and penetration state; define the acceleration signal characteristics of each state, and determine the cutting process based on the changes in vibration and frequency; Figure 2 As shown, specifically:

[0029] 201) In the cutting preparation state, the bone oscillating saw first accelerates to a predetermined speed and then approaches the experimental material to be cut. The acceleration signal in this state is mainly caused by the reciprocating motion of the bone oscillating saw itself. Figure 2 As shown, in this state, the amplitude of the frequency response is maintained at a low level and there is no record of exceeding 1.

[0030] 202) Pre-cutting state, marking the beginning of the cutting process. The bone saw first contacts the experimental material and gradually cuts into it. As the experimental material deforms and the bone saw goes deeper, the interaction between the two causes strong vibrations.

[0031] 203) Real-time cutting state is the core stage of the entire cutting process. During this period, the interaction between the cutting and the experimental material reaches its peak. Due to the strong interaction between the bone pendulum saw and the experimental material, the acceleration signal at this stage has a significant peak and oscillation mode, and the high-frequency component increases.

[0032] 204) critical state, which occurs when the bone saw is about to cut out of the experimental material and the cutting process is nearing its end. It is revealed that the acceleration signal at this stage shows a significant downward trend, although it is still accompanied by certain fluctuations, which indicates that the cutting activity has not yet been completely terminated.

[0033] 205) Penetration state, the bone saw is completely separated from the experimental material, and the cutting work is completed.

[0034] S3. Feature extraction: The sampling frequency of the acceleration signal is set to 4000 Hz, with 64 ms as a sampling frame, and the number of data points in each sampling frame is 256; after feature extraction, 15-dimensional features are obtained, including one-dimensional frequency domain features, five-dimensional wavelet features and nine-dimensional time domain features; specifically:

[0035] One-dimensional frequency domain features: A sampling frame is transformed by FFT to obtain its spectrum, and the frequency value corresponding to the maximum peak of the sampling frame in the frequency range of 200Hz to 300Hz is found as its fundamental frequency value. The fundamental frequency value is the selected one-dimensional frequency domain feature.

[0036] Five-dimensional wavelet features: Maximum overlap discrete wavelet transform (MODWT) is used to decompose the sampling frame; in this embodiment, a Symlet wavelet (sym4) with an order of 4 is used as a signal processing tool. Figure 3a and Figure 3b As shown, the sampling frame ( Figure 3a ) is decomposed into 5 layers (W1, W2, W3, W4, W5), and the frequency ranges of each layer are 1000-2000Hz, 500-1000Hz, 250-500Hz, 125-250Hz, and 60-125Hz. The signal energy corresponding to the decomposed (W1, W2, W3, W4, W5) layer is divided by the energy of the sampling frame signal to obtain 5 relative energies, which are the selected five-dimensional wavelet features.

[0037] Nine-dimensional time domain features: The W2 layer signal obtained above is introduced with the following statistical features: mean, variance, peak, effective value, peak factor, kurtosis, pulse factor, waveform factor, and margin factor. These 9 values ​​are the selected nine-dimensional time domain features.

[0038] S4. Based on the fifteen-dimensional features obtained by feature extraction, the acceleration signal in the cutting process is manually divided into the five cutting states defined in step S2, and the start and end points of each state are marked according to different feature changes; specifically, the state calibration process for each complete cutting process is as follows:

[0039] Calibration in the cutting preparation state: In a complete cutting process, the amplitude of the spectrum graph of the cutting preparation state does not exceed 1, so the initial moment until the amplitude exceeds 1 is marked as the cutting preparation state;

[0040] Pre-cutting state: the end of the cutting preparation marks the beginning of the pre-cutting state, and the end mark is that the fundamental frequency values ​​extracted from the three sampling frames are equal, marking the last sampling frame as the end of the pre-cutting state;

[0041] Real-time cutting state: the middle part between the end of pre-cutting state and the beginning of critical state;

[0042] Critical stage: the part of 3 sampling frames before the penetration state begins;

[0043] Penetration state: After the bone pendulum saw penetrates, the saw blade will block the optical path of the laser sensor and transmit a high-level signal to the main control chip, and the high-level signal is used to mark the penetration state;

[0044] S5. Establish a classification model; the classification model used in this embodiment is SVM (Support Vector Machine). The following are specific hyperparameter settings, BoxConstraint is set to 10. The value of GapTolerance is set to 0. This embodiment selects polynomial as the KernelFunction, that is, the polynomial kernel function. KernelScale is set to 10. PolynomialOrder is set to 2, which indicates the order of the polynomial kernel function, which determines the dimension of the data when it is mapped to a high-dimensional space, thereby affecting the nonlinear fitting ability of the model. The RemoveDuplicates parameter is set to true, which means that duplicate observations in the data set will be removed to reduce the impact of redundant information on model training and improve computational efficiency. The value of ShrinkagePeriod is set to 0, table. Finally, ISDA (Iterative Single Data Algorithm) was selected as the Solver.

[0045] The total number of samples is 1415. The dataset consists of a total of 5 different cutting states, each containing 283 samples. 80% of the dataset is used as a training set and 20% of the dataset is used as a test set to evaluate the performance of the classification network model, which means that the number of training sets and test sets are 1132 and 283 respectively, and the classification network model outputs 5 different cutting stages.

[0046] The confusion matrix of the test set is as follows Figure 4 As shown, the critical and penetration predictions achieve a 100% recognition rate.

[0047] Preferably, an application of a cutting state classification model for a medical bone oscillating saw when cutting human bones is based on the above-mentioned construction method of the cutting state classification model for a medical bone oscillating saw when cutting human bones and a cutting state recognition device, wherein the accelerometer of the cutting state recognition device is used to collect signals and calculate the fundamental frequency amplitude; feature extraction is performed on the collected signals and input into a support vector machine (SVM) classification model for classification; and the classification results are displayed in real time to assist doctors in judging the cutting state. Specific application methods are as follows: Figure 5 As shown, first, the STM32 main control chip initializes each sensor, and then starts manually. The accelerometer starts to collect acceleration signals. First, a frame of signal is collected for FFT calculation to determine whether its fundamental frequency amplitude is greater than 1. If it is less than or equal to 1, a frame of signal is collected again. If it is greater than 1, the frame data is sent to the computer through the serial port for 15-dimensional feature extraction, and then the features are sent to the pre-trained SVM classification model for classification, and then the classification results are displayed in real time.

[0048] Preferably, the embodiment of the present application also provides a specific implementation of an electronic device capable of implementing all steps in the method for constructing a cutting state classification model when a medical bone oscillating saw cuts human bones in the above embodiment, and the electronic device specifically includes the following contents:

[0049] Processor, memory, communications interface and bus;

[0050] Among them, the processor, memory, and communication interface communicate with each other through the bus; the communication interface is used to realize information transmission between related devices such as server-side devices, metering devices, and user-side devices.

[0051] The processor is used to call the computer program in the memory, and when the processor executes the computer program, all the steps in the method for constructing a cutting state classification model when the medical bone oscillating saw cuts human bones in the above embodiment are implemented.

[0052] An embodiment of the present application also provides a computer-readable storage medium that is capable of implementing all steps in the method for constructing a cutting state classification model when a medical bone oscillating saw cuts human bones in the above-mentioned embodiment. A computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, all steps in the method for constructing a cutting state classification model when a medical bone oscillating saw cuts human bones in the above-mentioned embodiment are implemented.

[0053] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the hardware + program embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0054] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0055] Although the present application provides method operation steps such as embodiments or flow charts, more or fewer operation steps may be included based on conventional or non-creative labor. The order of steps listed in the embodiments is only one way of executing the order of many steps and does not represent the only execution order. When the actual device or client product is executed, it can be executed in the order of the method shown in the embodiments or the drawings or in parallel (for example, in a parallel processor or multi-threaded processing environment).

[0056] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0057] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0058] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0059] The present invention is not limited to the embodiments described above. The above description of the specific embodiments is intended to describe and illustrate the technical solution of the present invention. The above specific embodiments are merely illustrative and not restrictive. Without departing from the scope of the present invention and the scope of protection of the claims, a person of ordinary skill in the art can also make many forms of specific changes under the guidance of the present invention, which all fall within the scope of protection of the present invention.

Claims

1. A method for constructing a cutting state classification model when a medical bone oscillating saw cuts human bones, characterized in that: include: S1. Prepare artificial cancellous bone as experimental material; S2. Classify the cutting states into cutting preparation state, pre-cutting state, real-time cutting state, critical state and penetration state; define the acceleration signal characteristics of each state, and determine the cutting process based on the changes in vibration and frequency; S3. Set the sampling frequency of the acceleration signal to 4000 Hz, with each 64 ms as a sampling frame, and obtain 15-dimensional features through feature extraction; Including one-dimensional frequency domain features, five-dimensional wavelet features and nine-dimensional time domain features; S4. Based on the fifteen-dimensional features obtained by feature extraction, the acceleration signal during the cutting process is manually divided into the five cutting states defined in step S2, and the start and end points of each state are marked according to different feature changes; S5. Establish a classification model; use the support vector machine SVM classification model and adopt a polynomial kernel function.

2. The method for constructing a cutting state classification model when the medical bone oscillating saw cuts human bones according to claim 1, characterized in that: The experimental material of the artificial cancellous bone in step S1 is solid hard polyurethane foam plastic with a thickness of 4 mm.

3. The method for constructing a cutting state classification model when the medical bone oscillating saw cuts human bones according to claim 1, characterized in that: In step S2, the definitions of each state are as follows: In the cutting preparation state, the bone oscillating saw first accelerates to a predetermined speed and then approaches the experimental material to be cut; the acceleration signal in this state is caused by the reciprocating motion of the bone oscillating saw itself; The pre-cutting state marks the beginning of the cutting process, when the bone pendulum saw first contacts the experimental material and gradually cuts into it; As the experimental material deforms and the bone pendulum saw penetrates deeper, the interaction between the two causes vibrations; Real-time cutting state, during which the interaction between the bone pendulum saw and the experimental material reaches its peak; the acceleration signal in this state has a significant peak and oscillation mode, and the high-frequency component increases; The critical state occurs when the bone saw is about to cut out of the experimental material and the cutting process is nearing the end; In this state, the acceleration signal shows a significant downward trend and is accompanied by fluctuations, indicating that the cutting activity has not been completely terminated; In the penetration state, the bone saw is completely separated from the experimental material and the cutting work is completed.

4. The method for constructing a cutting state classification model when the medical bone oscillating saw cuts human bones according to claim 1, characterized in that: In step S3, One-dimensional frequency domain features: a sampling frame is transformed by FFT to obtain its spectrum diagram, and the frequency value corresponding to the maximum peak value of the sampling frame is found in the frequency range of 200Hz to 300Hz as its fundamental frequency value; the fundamental frequency value is the selected one-dimensional frequency domain feature; Five-dimensional wavelet features, using maximum overlap discrete wavelet transform MODWT to decompose the sampling frame and extract 5 relative energies as five-dimensional wavelet features; The nine-dimensional time domain features include nine statistical features introduced into the sampling frame: mean, variance, peak, effective value, peak factor, kurtosis, pulse factor, waveform factor, and margin factor.

5. The method for constructing a cutting state classification model when a medical bone oscillating saw cuts human bones according to claim 4, characterized in that: A Symlet wavelet of order 4 is used as a signal processing tool; the sampling frame is decomposed into 5 layers W1, W2, W3, W4, and W5 using MODWT, and the frequency range of each layer is 1000-2000Hz, 500-1000Hz, 250-500Hz, 125-250Hz, and 60-125Hz respectively; the signal energy corresponding to the decomposed W1, W2, W3, W4, and W5 layers is divided by the energy of the sampling frame signal to obtain five relative energies; these five relative energies are the five-dimensional wavelet features.

6. The method for constructing a cutting state classification model when the medical bone oscillating saw cuts human bones according to claim 1, characterized in that: The process of calibrating the state of each complete cutting process in step S4 is as follows: Calibration in the cutting preparation state: In a complete cutting process, the amplitude of the spectrum graph of the cutting preparation state does not exceed 1, so the initial moment until the amplitude exceeds 1 is marked as the cutting preparation state; Pre-cutting state: the end of the cutting preparation marks the beginning of the pre-cutting state, and the end mark is that the fundamental frequency values ​​extracted from the three sampling frames are equal, marking the last sampling frame as the end of the pre-cutting state; Real-time cutting state: the middle part between the end of pre-cutting state and the beginning of critical state; Critical stage: the part of 3 sampling frames before the penetration state begins; Penetration status: After the bone pendulum saw penetrates, the saw blade will block the optical path of the laser sensor and transmit a high-level signal to the main control chip, using the high-level signal to mark the penetration status.

7. A device for identifying the cutting state of a medical bone oscillating saw when cutting human bones, used in the method for constructing a cutting state classification model of a medical bone oscillating saw when cutting human bones as described in any one of claims 1 to 6, characterized in that: Including main control chip, storage module, accelerometer, communication module, laser sensor, display module and power module; The main control chip is respectively connected to the storage module, the accelerometer, the communication module, the laser sensor, the display module and the power module; The communication module uses Bluetooth transmission, the display module uses a serial port screen, the laser sensor is used to calibrate the penetration state, and the detection frequency is 100 Hz; the cutting direction of the bone pendulum saw is to cut downward perpendicular to the bone plate, and the collected acceleration signal is the vibration direction of the bone pendulum saw.

8. An application of a cutting state classification model when a medical bone oscillating saw cuts human bones, based on the method for constructing a cutting state classification model when a medical bone oscillating saw cuts human bones as described in any one of claims 1 to 6 and the cutting state recognition device of claim 7, characterized in that: The accelerometer of the cutting state recognition device collects signals and calculates the fundamental frequency amplitude; the collected signals are feature extracted and input into the support vector machine (SVM) classification model for classification; the classification results are displayed in real time to assist doctors in judging the cutting state.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the method for constructing a cutting state classification model when the medical bone oscillating saw cuts human bones are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for constructing a cutting state classification model when a medical bone oscillating saw cuts human bones as described in any one of claims 1 to 6 are implemented.

Images