Monitoring device and monitoring method thereof

Abstract

The present invention relates to a monitoring method, comprising: using a plurality of sensors to collect physiological parameter signals of a patient; converting the physiological parameter signals into digital signals; issuing an indicator exceeding alarm when any one of the physiological parameters exceeds a preset range; using drip images as samples, inputting them into a deep learning model for training to obtain an infusion progress estimation model; using the infusion progress estimation model to monitor the hanging bottle of a target patient, and issuing an alarm to replace the hanging bottle when the liquid in the hanging bottle is less than a preset value. The present invention can realize synchronous monitoring of multiple vital signs such as heart rate, blood pressure, and oxygen saturation by using a plurality of sensors to simultaneously collect physiological parameter signals of a patient. When any one of the physiological parameters exceeds a preset range, an alarm is automatically issued, which can promptly remind medical staff to pay attention to the abnormal state of the patient, reduce the risk of missed monitoring and misjudgment, and improve the quality of nursing.

Description

Technical Field

[0001] The present invention relates to the technical field of clinical nursing, and in particular to a monitoring device and a monitoring method thereof. Background Art

[0002] With the continuous development of medical technology and the transformation of medical models, the demand for monitoring in clinical care is increasing. Especially in environments such as intensive care units (ICUs) and operating rooms where vital signs monitoring is extremely demanding, timely and accurate monitoring of patients' physiological parameters has become a key factor in improving the quality of medical services. At present, common monitoring equipment mainly includes ECG monitors, blood pressure monitors, pulse oximeters, etc., which can monitor patients' heart rate, blood pressure, oxygen saturation and other basic vital signs in real time. However, these traditional devices still have shortcomings in many aspects.

[0003] First, many monitoring devices do not provide adequate alarm mechanisms. When the patient's condition changes dramatically, medical staff may not be able to detect it in time. In addition, existing monitoring devices can usually only monitor one or several vital signs, lacking comprehensive and multi-parameter synchronous monitoring capabilities. This requires medical staff to use multiple devices during the actual care process, increasing the complexity of the operation and the workload. Summary of the invention

[0004] To solve the above problems, an object of the embodiments of the present invention is to provide a monitoring device and a monitoring method thereof.

[0005] A monitoring method, comprising:

[0006] Step 1: Use a variety of sensors to collect the patient's physiological parameter signals;

[0007] Step 2: Convert the physiological parameter signal into a digital signal;

[0008] Step 3: Preprocessing the digital signal to form real-time physiological parameter information of the patient, and displaying the real-time physiological parameter information;

[0009] Step 4: When any physiological parameter exceeds the preset range, an indicator exceeding alarm is issued;

[0010] Step 5: Obtain the drip images of the patient during infusion;

[0011] Step 6: Use the drip images as samples and input them into the deep learning model for training to obtain an infusion progress estimation model;

[0012] Step 7: Use the infusion progress estimation model to monitor the target patient's IV bottle, and when the liquid in the IV bottle is less than a preset value, issue an alarm to replace the IV bottle.

[0013] Preferably, the step 3: preprocessing the digital signal to form real-time physiological parameter information of the patient, and displaying the real-time physiological parameter information, comprises:

[0014] Step 3.1: Use Sym wavelet, db wavelet and coif wavelet to decompose the digital signal and obtain the wavelet coefficients at the corresponding decomposition scale;

[0015] Step 3.2: Determine the threshold value according to the wavelet coefficient;

[0016] Step 3.3: When the wavelet coefficient is less than the threshold, the wavelet coefficient is changed to 0;

[0017] Step 3.4: When the wavelet coefficient is greater than the threshold, the threshold processing model is used to process the wavelet coefficient to form a pre-processed wavelet coefficient, and the pre-processed wavelet coefficient is reconstructed to form a reconstructed digital signal;

[0018] Step 3.5: Generating real-time physiological parameter information of the patient based on the reconstructed digital signal.

[0019] Preferably, in step 3.4, the threshold processing model is:

[0020]

[0021] in, represents the wavelet coefficients after preprocessing, W j,k represents the original wavelet coefficients, λ represents the threshold, a represents the adjustment coefficient, and sgn() represents the sign function.

[0022] Preferably, the step 3.5: forming the patient's real-time physiological parameter information based on the reconstructed digital signal comprises:

[0023] Step 3.5.1: Evaluate the signal quality of the reconstructed digital signal to form an evaluation value; wherein the calculation formula of the evaluation value is:

[0024]

[0025] Among them, s(n) represents the original digital signal, represents the reconstructed digital signal, n represents the sampling point number, N represents the sampling length, SNR represents the first evaluation value, and RMSE represents the second evaluation value;

[0026] Step 3.5.2: Extract the digital signal whose evaluation value is within the set range as the final reconstructed digital signal.

[0027] Preferably, the step 6: taking the drip images as samples and inputting them into a deep learning model for training to obtain an infusion progress estimation model comprises:

[0028] Step 6.1: Take a sliding window centered at each pixel of the drip image and calculate the intensity of the noise point under the sliding window;

[0029] Step 6.2: When the intensity of the noise point is greater than a preset threshold, the drip image in the corresponding sliding window is smoothed to obtain a smoothed drip image;

[0030] Step 6.3: continuously moving the sliding window until the smoothing process of the entire drip image is completed to obtain a smoothed drip image;

[0031] Step 6.4: Perform equalization processing on the smoothed drip image to obtain a sample;

[0032] Step 6.5: Input the sample into the deep learning model for training to obtain an infusion progress estimation model.

[0033] Preferably, in step 6.1, the intensity of the noise point under the sliding window is:

[0034]

[0035] Among them, I(i) represents the pixel value of the pixel in the sliding window, N represents the number of pixels in the sliding window, μ represents the mean value of the pixel in the sliding window, and σ represents the intensity of the noise under the sliding window.

[0036] Preferably, in step 6.3, each pixel point in the sliding window is collected to form a pixel set, and the mode of the pixel set is taken as the output of the pixel point after the sliding window is smoothed; wherein the output formula of the pixel point after smoothing is:

[0037] I denoised (x,y)=Mode(I original (xh,yh),…,I original (x+h,y+h))

[0038] In the formula, I denoised (x, y) is the output of the smoothed pixel, h is the side length of the sliding window, I original (x, y) is the pixel value of the pixel at the (x, y) position on the sliding window.

[0039] Preferably, the step 6.5: inputting the sample into a deep learning model for training to obtain an infusion progress estimation model comprises:

[0040] Step 6.5.1: Set a label for each sample to indicate the infusion progress;

[0041] Step 6.5.2: Input the sample into the VGG network for training;

[0042] Step 6.5.3: Use the loss function to evaluate the training effect of the VGG network. When the value of the loss function reaches a preset value, the training is completed and the infusion progress estimation model is obtained; wherein the loss function is:

[0043]

[0044] in, is, C is the number of infusion progress labels, y i The actual infusion progress. Labels predicted by the VGG network.

[0045] The present invention also provides a monitoring device, comprising:

[0046] A physiological parameter signal acquisition module is used to acquire physiological parameter signals of patients using a variety of sensors;

[0047] A digital signal conversion module, used for converting the physiological parameter signal into a digital signal;

[0048] A signal preprocessing module, used for preprocessing the digital signal to form real-time physiological parameter information of the patient, and displaying the real-time physiological parameter information;

[0049] The indicator exceeding module is used to issue an indicator exceeding alarm when any physiological parameter exceeds the preset range;

[0050] A drip image acquisition module is used to acquire drip images of patients during infusion;

[0051] A training module is used to input the drip images as samples into a deep learning model for training to obtain an infusion progress estimation model;

[0052] The monitoring module is used to monitor the target patient's IV bottle using the infusion progress estimation model, and to issue an alarm to replace the IV bottle when the liquid in the IV bottle is less than a preset value.

[0053] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of the above-mentioned monitoring method when executed by a processor.

[0054] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0055] The present invention relates to a monitoring method. Compared with the prior art, the present invention can realize synchronous monitoring of multiple vital signs such as heart rate, blood pressure, oxygen saturation, etc. by using multiple sensors to simultaneously collect physiological parameter signals of patients. When any physiological parameter exceeds a preset range, an alarm is automatically issued, which can promptly remind medical staff to pay attention to the abnormal state of the patient, reduce the risk of missed monitoring and misjudgment, and improve the quality of care. In addition, the present invention can realize intelligent monitoring of the progress of infusion by acquiring drip images during infusion and using a deep learning model for training and estimation. This not only improves the automation level of infusion management, but also reduces the time and energy consumption of medical staff in observing the infusion status.

[0056] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0058] Figure 1 A flow chart of a monitoring method in an embodiment provided by the present invention;

[0059] Figure 2 A schematic diagram of denoising effects of different wavelet bases in the embodiments provided by the present invention;

[0060] Figure 3 This is a schematic diagram of a monitoring device in an embodiment of the present invention. DETAILED DESCRIPTION

[0061] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0062] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0063] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0064] See also Figure 1 , a monitoring method, comprising:

[0065] Step 1: Use a variety of sensors to collect the patient's physiological parameter signals;

[0066] Step 2: Convert the physiological parameter signal into a digital signal;

[0067] Step 3: Preprocess the digital signal to form the patient's real-time physiological parameter information, and display the real-time physiological parameter information; the physiological parameter information includes: heart rate, blood pressure, oxygen saturation, body temperature and other vital signs.

[0068] Further, step 3: pre-processing the digital signal to form real-time physiological parameter information of the patient, and displaying the real-time physiological parameter information, including:

[0069] Step 3.1: Use Sym wavelet, db wavelet and coif wavelet to decompose the digital signal to obtain wavelet coefficients under the corresponding decomposition scale; the present invention uses three wavelet bases to test and obtain the improvement graph of the signal-to-noise ratio of the three wavelets as shown in the figure. Figure 2 As shown. Figure 2 It can be seen that the present invention uses the Sym6 wavelet for decomposition at its best.

[0070] Step 3.2: Determine the threshold value according to the wavelet coefficient;

[0071] Step 3.3: When the wavelet coefficient is less than the threshold, the wavelet coefficient is changed to 0;

[0072] Step 3.4: When the wavelet coefficient is greater than the threshold, the wavelet coefficient is processed using the threshold processing model to form a pre-processed wavelet coefficient, and the pre-processed wavelet coefficient is reconstructed to form a reconstructed digital signal; wherein the threshold processing model is:

[0073]

[0074] in, represents the wavelet coefficients after preprocessing, W j,k represents the original wavelet coefficients, λ represents the threshold, a represents the adjustment coefficient, and sgn() represents the sign function.

[0075] Step 3.5: Generating real-time physiological parameter information of the patient based on the reconstructed digital signal;

[0076] Furthermore, step 3.5 includes:

[0077] Step 3.5.1: Evaluate the signal quality of the reconstructed digital signal to form an evaluation value; wherein the calculation formula of the evaluation value is:

[0078]

[0079] Among them, s(n) represents the original digital signal, represents the reconstructed digital signal, n represents the sampling point number, N represents the sampling length, SNR represents the first evaluation value, and RMSE represents the second evaluation value;

[0080] Step 3.5.2: Extract the digital signal whose evaluation value is within the set range as the final reconstructed digital signal.

[0081] The present invention can effectively remove noise signals by performing wavelet transform processing on digital signals, thereby improving the quality of digital signals and thus improving the accuracy and reliability of physiological parameter information.

[0082] Step 4: When any physiological parameter exceeds the preset range, an indicator exceeding alarm is issued;

[0083] Step 5: Obtain the drip images of the patient during infusion;

[0084] Step 6: Use the drip images as samples and input them into the deep learning model for training to obtain an infusion progress estimation model;

[0085] Further, step 6: the drip images are used as samples and input into the deep learning model for training to obtain an infusion progress estimation model, including:

[0086] Step 6.1: Take a sliding window centered at each pixel of the drip image and calculate the intensity of the noise point under the sliding window;

[0087] In step 6.1, the intensity of the noise point under the sliding window is:

[0088]

[0089] Among them, I(i) represents the pixel value of the pixel in the sliding window, N represents the number of pixels in the sliding window, μ represents the mean value of the pixel in the sliding window, and σ represents the intensity of the noise under the sliding window.

[0090] Step 6.2: When the intensity of the noise point is greater than a preset threshold, the drip image in the corresponding sliding window is smoothed to obtain a smoothed drip image;

[0091] Step 6.3: continuously moving the sliding window until the smoothing process of the entire drip image is completed to obtain a smoothed drip image;

[0092] In step 6.3, each pixel point in the sliding window is collected to form a pixel set, and the mode of the pixel set is taken as the output of the pixel point after the sliding window is smoothed; wherein the output formula of the pixel point after smoothing is:

[0093] I denoised (x,y)=Mode(I original (xh,yh),…,I original (x+h,y+h))

[0094] In the formula, I denoised (x, y) is the output of the smoothed pixel, h is the side length of the sliding window, I original (x, y) is the pixel value of the pixel at the (x, y) position on the sliding window.

[0095] Step 6.4: Perform equalization processing on the smoothed droplet image to obtain a sample; in the embodiment of the present invention, the image may be processed by histogram equalization to obtain the sample.

[0096] The present invention can effectively reduce random noise in the image by smoothing the drip image, so that the neural network can focus more on important features during training instead of being disturbed by noise; the contrast of the drip image can be enhanced by histogram equalization, making the important features in the image more obvious. This helps the neural network to better learn and extract key features.

[0097] Step 6.5: inputting the sample into a deep learning model for training to obtain an infusion progress estimation model;

[0098] Furthermore, step 6.5 includes:

[0099] Step 6.5.1: Set a label for each sample to indicate the infusion progress;

[0100] Step 6.5.2: Input the sample into the VGG network for training;

[0101] Step 6.5.3: Use the loss function to evaluate the training effect of the VGG network. When the value of the loss function reaches a preset value, the training is completed and the infusion progress estimation model is obtained; wherein the loss function is:

[0102]

[0103] in, is, C is the number of infusion progress labels, y i The actual infusion progress. Labels predicted by the VGG network.

[0104] Step 7: Use the infusion progress estimation model to monitor the target patient's IV bottle, and when the liquid in the IV bottle is less than a preset value, issue an alarm to replace the IV bottle.

[0105] The present invention can realize synchronous monitoring of multiple vital signs such as heart rate, blood pressure, oxygen saturation, etc. by using multiple sensors to simultaneously collect physiological parameter signals of patients. When any physiological parameter exceeds the preset range, an alarm is automatically issued, which can promptly remind medical staff to pay attention to the abnormal state of the patient, reduce the risk of missed monitoring and misjudgment, and improve the quality of care. In addition, the present invention can realize intelligent monitoring of the progress of infusion by acquiring drip images during infusion and using deep learning models for training and estimation. This not only improves the level of automation of infusion management, but also reduces the time and energy consumption of medical staff in observing the infusion status.

[0106] See also Figure 3 The present invention also provides a monitoring device, comprising:

[0107] A physiological parameter signal acquisition module is used to acquire physiological parameter signals of patients using a variety of sensors;

[0108] A digital signal conversion module, used for converting the physiological parameter signal into a digital signal;

[0109] A signal preprocessing module, used for preprocessing the digital signal to form real-time physiological parameter information of the patient, and displaying the real-time physiological parameter information;

[0110] The indicator exceeding module is used to issue an indicator exceeding alarm when any physiological parameter exceeds the preset range;

[0111] A drip image acquisition module is used to acquire drip images of patients during infusion;

[0112] A training module is used to input the drip images as samples into a deep learning model for training to obtain an infusion progress estimation model;

[0113] The monitoring module is used to monitor the target patient's IV bottle using the infusion progress estimation model, and to issue an alarm to replace the IV bottle when the liquid in the IV bottle is less than a preset value.

[0114] Compared with the prior art, the beneficial effects of the monitoring device provided by the present invention are the same as the beneficial effects of the monitoring method described in the above technical solution, which will not be elaborated here.

[0115] The present invention also provides 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 in the above-mentioned monitoring method are implemented. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present invention are the same as the beneficial effects of the monitoring method described in the above-mentioned technical solution, which will not be repeated here.

[0116] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technical solution that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A monitoring method, characterized in that: include: Step 1: Acquire patient physiological parameter signals collected by multiple sensors; Step 2: Convert the physiological parameter signal into a digital signal; Step 3: Preprocessing the digital signal to form real-time physiological parameter information of the patient, and displaying the real-time physiological parameter information; Step 4: When any physiological parameter exceeds the preset range, an indicator exceeding alarm is issued; Step 5: Obtain the drip images of the patient during infusion; Step 6: Use the drip images as samples and input them into the deep learning model for training to obtain an infusion progress estimation model; Step 6: using the drip image as a sample and inputting it into a deep learning model for training to obtain an infusion progress estimation model, including: Step 6.1: Take a sliding window centered at each pixel of the drip image and calculate the intensity of the noise point under the sliding window; In step 6.1, the intensity of the noise point under the sliding window is: Where I(i) represents the pixel value of the pixel in the sliding window, N represents the number of pixels in the sliding window, μ represents the mean value of the pixel in the sliding window, and σ represents the intensity of the noise under the sliding window; Step 6.2: When the intensity of the noise point is greater than a preset threshold, the drip image in the corresponding sliding window is smoothed to obtain a smoothed drip image; Step 6.3: continuously moving the sliding window until the smoothing process of the entire drip image is completed to obtain a smoothed drip image; In step 6.3, each pixel point in the sliding window is collected to form a pixel set, and the mode of the pixel set is taken as the output of the pixel point after the sliding window is smoothed; wherein the output formula of the pixel point after smoothing is: I denoised (x,y)=Mode(I original (x-h,y-h),…,I original (x+h,y+h)) In the formula, I denoised (x, y) is the output of the smoothed pixel, h is the side length of the sliding window, I original (x, y) is the pixel value of the pixel point at the position (x, y) on the sliding window, and Mode() means taking the mode; Step 6.4: Perform equalization processing on the smoothed drip image to obtain a sample; Step 6.5: inputting the sample into a deep learning model for training to obtain an infusion progress estimation model; The step 6.5: inputting the sample into a deep learning model for training to obtain an infusion progress estimation model, including: Step 6.5.1: Set a label for each sample to indicate the infusion progress; Step 6.5.2: Input the sample into the VGG network for training; Step 6.5.3: Use the loss function to evaluate the training effect of the VGG network. When the value of the loss function reaches a preset value, the training is completed and the infusion progress estimation model is obtained; wherein the loss function is: in, is the value of the loss function, C is the number of infusion progress labels, y i The actual infusion progress. Labels predicted by the VGG network; Step 7: Use the infusion progress estimation model to monitor the target patient's IV bottle, and when the liquid in the IV bottle is less than a preset value, issue an alarm to replace the IV bottle.

2. A monitoring method according to claim 1, characterized in that: The step 3: pre-processing the digital signal to form the patient's real-time physiological parameter information, and displaying the real-time physiological parameter information, includes: Step 3.1: Use Sym wavelet, db wavelet and coif wavelet to decompose the digital signal and obtain the wavelet coefficients at the corresponding decomposition scale; Step 3.2: Determine the threshold value according to the wavelet coefficient; Step 3.3: When the wavelet coefficient is less than the threshold, the wavelet coefficient is changed to 0; Step 3.4: When the wavelet coefficient is greater than the threshold, the threshold processing model is used to process the wavelet coefficient to form a pre-processed wavelet coefficient, and the pre-processed wavelet coefficient is reconstructed to form a reconstructed digital signal; Step 3.5: Generating real-time physiological parameter information of the patient based on the reconstructed digital signal.

3. A monitoring method according to claim 2, characterized in that: In step 3.4, the threshold processing model is: in, represents the wavelet coefficients after preprocessing, W j,k represents the original wavelet coefficients, λ represents the threshold, a represents the adjustment coefficient, and sgn() represents the sign function.

4. A monitoring method according to claim 3, characterized in that: The step 3.5: forming the patient's real-time physiological parameter information based on the reconstructed digital signal, comprises: Step 3.5.1: Evaluate the signal quality of the reconstructed digital signal to form an evaluation value; Step 3.5.2: Extract the digital signal whose evaluation value is within the set range as the final reconstructed digital signal.

5. A monitoring device, characterized in that: include: A physiological parameter signal acquisition module is used to obtain patient physiological parameter signals collected by various sensors; A digital signal conversion module, used for converting the physiological parameter signal into a digital signal; A signal preprocessing module, used for preprocessing the digital signal to form real-time physiological parameter information of the patient, and displaying the real-time physiological parameter information; The indicator exceeding module is used to issue an indicator exceeding alarm when any physiological parameter exceeds the preset range; A drip image acquisition module is used to acquire drip images of patients during infusion; A training module is used to input the drip images as samples into a deep learning model for training to obtain an infusion progress estimation model; The drip images are used as samples and input into the deep learning model for training to obtain the infusion progress estimation model, including: Step 6.1: Take a sliding window centered at each pixel of the drip image and calculate the intensity of the noise point under the sliding window; In step 6.1, the intensity of the noise point under the sliding window is: Where I(i) represents the pixel value of the pixel in the sliding window, N represents the number of pixels in the sliding window, μ represents the mean value of the pixel in the sliding window, and σ represents the intensity of the noise under the sliding window; Step 6.2: When the intensity of the noise point is greater than a preset threshold, the drip image in the corresponding sliding window is smoothed to obtain a smoothed drip image; Step 6.3: continuously moving the sliding window until the smoothing process of the entire drip image is completed to obtain a smoothed drip image; In step 6.3, each pixel point in the sliding window is collected to form a pixel set, and the mode of the pixel set is taken as the output of the pixel point after the sliding window is smoothed; wherein the output formula of the pixel point after smoothing is: I denoised (x,y)=Mode(I original (x-h,y-h),…,I original (x+h,y+h)) In the formula, I denoised (x, y) is the output of the smoothed pixel, h is the side length of the sliding window, I original (x, y) is the pixel value of the pixel point at the position (x, y) on the sliding window, and Mode() means taking the mode; Step 6.4: Perform equalization processing on the smoothed drip image to obtain a sample; Step 6.5: inputting the sample into a deep learning model for training to obtain an infusion progress estimation model; The step 6.5: inputting the sample into a deep learning model for training to obtain an infusion progress estimation model, including: Step 6.5.1: Set a label for each sample to indicate the infusion progress; Step 6.5.2: Input the sample into the VGG network for training; Step 6.5.3: Use the loss function to evaluate the training effect of the VGG network. When the value of the loss function reaches a preset value, the training is completed and the infusion progress estimation model is obtained; wherein the loss function is: in, is the value of the loss function, C is the number of infusion progress labels, y i The actual infusion progress. Labels predicted by the VGG network; The monitoring module is used to monitor the target patient's IV bottle using the infusion progress estimation model, and to issue an alarm to replace the IV bottle when the liquid in the IV bottle is less than a preset value.

6. 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 in a monitoring method according to any one of claims 1 to 4 are implemented.

Images