Silent infusion monitoring method, system, storage medium and equipment
Through the Transformer algorithm, the infusion bottle model is identified and the infusion progress is calculated, silent infusion monitoring is achieved, and the problems of high noise and low intelligence in the existing technology are solved, improving the patient's medical experience and the work efficiency of medical workers.
Patent Information
- Application Number
- CN202411254877.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-09
- Publication Date
- 2024-12-03
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing infusion monitoring system is noisy and has low intelligence, which makes it easy for medical workers to set alarm thresholds when they are busy, affecting the safety and experience of patients.
The Transformer algorithm is used to build a model identification model. By collecting the weight and image data of the infusion bottle/infusion bag, the model is automatically identified and the infusion progress is calculated, and a reminder is issued in a silent manner to reduce the workload of medical workers and create a silent medical treatment environment.
It improves the automation of infusion monitoring, reduces the workload of medical workers, ensures the safety of patients and medical treatment experience, and avoids the anxiety caused by noise.
Smart Images

Figure CN119055881A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of infusion monitoring, and in particular relates to a silent infusion monitoring method, system, storage medium and equipment. Background Art
[0002] Infusion is a commonly used treatment method, which requires the patient or the accompanying person to notify the nurse to change the dressing or remove the needle. If the notification is not timely, dangerous situations such as blood reflux or air input into the vein may occur. However, in the actual infusion process, some elderly patients seek medical treatment alone, which is prone to untimely notification. In addition, some hospitals will arrange medical workers to conduct inspections and carry out the next step in time, but when the nurse's workload is too heavy, there is a high probability that the patient will not change the dressing in time, resulting in dangerous situations. This will increase the workload of nurses, and it is difficult to ensure the safety of patients, affecting the treatment effect of patients.
[0003] In the prior art, the infusion monitoring system based on the gravity weighing sensor monitors the weight of the infusion bottle / infusion bag through a pressure sensor to monitor the infusion process. When the infusion is about to be completed, an alarm is sounded through a buzzer to prompt the medical worker to change the dressing or remove the needle; however, in actual infusion treatment, due to different conditions of patients, the models of patients' infusion bottles / infusion bags are also different, so different alarm thresholds are required. Once the medical workers are busy, it is very easy to set the alarm threshold incorrectly or forget to set it, which will cause the notification to be advanced or delayed, making it difficult to ensure the safety of the patient's visit, affecting the patient's treatment effect and experience. At the same time, the buzzer noise is relatively large, and this sound may make the patient feel anxious, affecting the patient's treatment effect. Summary of the invention
[0004] The purpose of the present invention is to provide a silent infusion monitoring method, system, storage medium and device to solve the problems of high noise and low intelligence of the existing infusion monitoring system.
[0005] The present invention provides the following technical solutions:
[0006] In a first aspect, a silent infusion monitoring method is provided, comprising:
[0007] Collect weight data of infusion bottles / infusion bags at various locations in the infusion room;
[0008] Acquire an infusion image of the patient's location;
[0009] Based on the infusion image of the patient's location, the constructed model recognition model is used to identify the model of the infusion bottle / infusion bag of the patient at the current location;
[0010] Determine the empty weight of the infusion bottle / infusion bag at the patient's location based on the preset infusion bottle / infusion bag weight that matches the model and the identified infusion bottle / infusion bag model;
[0011] Based on the empty weight of the infusion bottle / infusion bag at each patient's location and the collected weight data of the infusion bottle / infusion bag at each patient's location, the infusion progress of the current patient is obtained;
[0012] The infusion progress of the current patient is displayed on the display terminal of the medical worker, and when the infusion progress of the current patient is greater than the set threshold, a reminder is issued in a silent manner.
[0013] Optionally, the collecting of the infusion image of the patient's location includes:
[0014] The current weight data of the collected infusion bottles / infusion bags at each location are compared with the set minimum threshold. If the current weight data of the infusion bottle / infusion bag at the current location is greater than the set minimum threshold, it indicates that there is a patient at the current location, and the infusion image is collected for the patient's location. Otherwise, there is no patient at the current location, and the infusion image is not collected for the current location.
[0015] Optionally, the model identification model is constructed using a Transformer algorithm;
[0016] The infusion image based on the patient's location, using the constructed model recognition model to identify the model of the infusion bottle / infusion bag of the patient at the current location, includes:
[0017] Step 3.1: The collected infusion image at the patient's location is sequentially divided into blocks and flattened to obtain an image patch vector;
[0018] Step 3.2: Input all image patch vectors into the current Transformer Block stacking layer for feature extraction to obtain feature blocks containing feature information;
[0019] Step 3.3: Perform Patch Merging on the feature map corresponding to the feature block, merge adjacent feature blocks, and input the merged feature block into the next Transformer Block stacking layer for feature extraction;
[0020] Step 3.4: Repeat step 3.3 several times until all feature information is extracted;
[0021] Step 3.5: Perform one-dimensional adaptive average pooling on all extracted feature information to obtain a global feature vector; and based on the global feature vector, classify through the Softmax classifier to achieve model recognition of the infusion bottle / infusion bag.
[0022] Optionally, the step 3.1 includes:
[0023] Step 3.1.1: preprocessing the acquired infusion image at the patient's location; the preprocessing includes cutting, brightening and contrast enhancement;
[0024] Step 3.1.2: Input the preprocessed infusion image into the Patch Partition module for partitioning;
[0025] Step 3.1.3: For each block after the image is divided, flatten it in the channel direction to obtain the image patch vector.
[0026] Optionally, in step 3.2, before all image patch vectors are input into the current Transformer Block stacking layer, the dimensions of the image patch vectors are adjusted through a linear Embedding layer; the Transformer Block stacking layer includes a window multi-head self-attention layer and a shifted window multi-head self-attention layer.
[0027] Optionally, in the step of obtaining the infusion progress of the current patient based on the empty weight of the infusion bottle / infusion bag at each patient's location and the collected weight data of the infusion bottle / infusion bag at each patient's location,
[0028] The infusion progress P of the current patient is:
[0029]
[0030] Among them, W in is the initial weight data of the infusion bottle / infusion bag at the patient’s location, W c The current weight data of the infusion bottle / infusion bag at the patient’s location, W e The empty weight of the infusion bottle / bag at the patient's location;
[0031] The infusion progress of the current patient is displayed on the display terminal of the medical worker, and when the infusion progress of the current patient is greater than the set threshold, a reminder is issued in a silent manner.
[0032] The display terminal includes a PDA of a medical worker and a central control screen of an infusion room.
[0033] In a second aspect, a silent infusion monitoring system is provided, comprising:
[0034] Weighing module, used to collect weight data of infusion bottles / infusion bags at various locations in the infusion room;
[0035] An image acquisition module, used to acquire an infusion image of the patient's location;
[0036] A model recognition module is used to identify the model of the infusion bottle / infusion bag of the patient at the current location based on the infusion image of the patient's location and using the constructed model recognition model;
[0037] An empty weight determination module, for determining the empty weight of the infusion bottle / infusion bag at the patient's location based on a preset infusion bottle / infusion bag weight matching the model and the identified infusion bottle / infusion bag model;
[0038] An infusion progress determination module, used to obtain the infusion progress of the current patient based on the empty weight of the infusion bottle / infusion bag at each patient's location and the collected weight data of the infusion bottle / infusion bag at each patient's location;
[0039] The display module is used to display the infusion progress of the current patient on the display terminal of the medical worker, and to issue a reminder in a silent manner when the infusion progress of the current patient is greater than a set threshold.
[0040] In a third aspect, a computer device is provided, comprising a processor and a memory; wherein, when the processor executes a computer program stored in the memory, the steps of the silent infusion monitoring method described in any one of the first aspects are implemented.
[0041] In a fourth aspect, a computer-readable storage medium is provided for storing a computer program; when the computer program is executed by a processor, the steps of the silent infusion monitoring method described in any one of the first aspects are implemented.
[0042] Compared with the prior art, the present invention has the following beneficial effects:
[0043] The present invention automatically identifies the model of the infusion bottle / infusion bag through a model recognition model, thereby effectively reducing the workload of medical workers. At the same time, it can automatically obtain the infusion progress of the current patient according to the collected weight of the infusion bottle / infusion bag, thereby facilitating the medical staff to monitor the infusion progress, and has a high degree of automation. In addition, the monitoring of the infusion by the present invention is in a silent state, which is conducive to creating a silent medical environment and improving the patient's medical experience. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] Figure 1 It is a flow chart of the overall steps of the silent infusion monitoring method of the present invention;
[0045] Figure 2 It is a schematic diagram of the installation position of the weighing module of the present invention;
[0046] Figure 3 It is a flow chart of constructing the model identification model of the present invention;
[0047] Figure 4 It is a structural block diagram of the silent infusion monitoring system of the present invention.
[0048] Marked in the figure: 1 is a weighing module, 2 is an intravenous infusion stand. DETAILED DESCRIPTION
[0049] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.
[0050] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0051] like Figure 1 As shown, a silent infusion monitoring method comprises the following steps:
[0052] Step 1: Collect weight data of infusion bottles / infusion bags at various locations in the infusion room.
[0053] Specifically, each position of the infusion room is provided with an intravenous infusion stand 2, and a weighing module 1 is provided on the intravenous infusion stand 2 to obtain the weight data of the infusion bottle / infusion bag, such as Figure 2 As shown, the collected weight data corresponds to each intravenous infusion stand 2 .
[0054] Step 2: Acquire an infusion image of the patient's location.
[0055] Specifically, by capturing the patient's infusion screen with an industrial camera, the captured image can be segmented to extract the image of the infusion bottle / infusion bag, and then the cut and extracted image can be preprocessed to enhance the edge features of the infusion bottle / infusion bag.
[0056] In this embodiment, the collected weight data of the infusion bottle / infusion bag at each position is compared with the set minimum threshold value. If the current weight data of the infusion bottle / infusion bag at the current position is greater than the set minimum threshold value, it indicates that there is a patient at the current position, and the infusion image is collected for the patient's position; otherwise, there is no patient at the current position, and the infusion image is not collected for the current position.
[0057] As an optional option, when the collected weight of the infusion bottle / infusion bag is less than 20N, there is no patient at the current location. At this time, the system enters the waiting state and does not collect the infusion image at the current location. When the collected weight of the infusion bottle / infusion bag exceeds 20N for the first time, it indicates that there is a patient at the current location, and the infusion image is collected for the patient's location.
[0058] Step 3: Based on the infusion image of the patient's location, use the constructed model recognition model to identify the infusion bottle / infusion bag model of the patient at the current location.
[0059] In this embodiment, the model identification model is constructed using the Transformer algorithm; Figure 3 As shown, when building the model recognition model, the test set and training set are constructed for the preprocessed images at a ratio of 30% and 70%, and training and optimization are continuously performed to obtain the optimal model recognition model.
[0060] As a further step, step 3 specifically includes:
[0061] Step 3.1: The collected infusion image at the patient's location is sequentially divided into blocks and flattened to obtain an image patch vector.
[0062] Specifically, step 3.1 includes:
[0063] Step 3.1.1: Preprocess the acquired infusion image at the patient's location; the preprocessing includes cutting, brightening and contrast enhancement.
[0064] Step 3.1.2: Input the preprocessed infusion image into the Patch Partition module for partitioning; as an option, partition the image into 4*4 pixel blocks.
[0065] Step 3.1.3: For each block after the image is divided, flatten it in the channel direction to obtain the image patch vector.
[0066] Step 3.2: Input all image patch vectors into the current Transformer Block stacking layer for feature extraction to obtain feature blocks containing feature information.
[0067] Before all image patch vectors are input into the current Transformer Block stacking layer, the dimension of the image patch vector is adjusted through the Linear Embedding layer; that is, the dimension of the image patch is changed to a preset value that can be accepted by the Transformer.
[0068] The Transformer Block stacked layers include a window multi-head self-attention layer and a shifted window multi-head self-attention layer.
[0069] Step 3.3: Perform the Patch Merging operation on the feature map corresponding to the feature block, merge adjacent feature blocks, and input the merged feature block into the next Transformer Block stacking layer for feature extraction.
[0070] Step 3.4: Repeat step 3.3 several times until all feature information is extracted.
[0071] Steps 3.3 and 3.4 are mainly used to extract multi-scale feature information and build a hierarchical Transformer model. The size is continuously increased through the Patch Merging operation, that is, the (H / 4)*(W / 4)*C feature map is divided into 4 feature maps with adjacent pixels of (H / 8)*(W / 8), and then spliced in the depth direction. The depth changes from C to 4C, so that multi-scale information can be obtained. Next, multiple stages are connected for processing. PatchMerging is performed on each stage first, and then some Transformer Blocks are used to reduce the dimension.
[0072] Step 3.5: Perform one-dimensional adaptive average pooling on all extracted feature information to obtain a global feature vector; and based on the global feature vector, classify through the Softmax classifier to achieve model recognition of the infusion bottle / infusion bag.
[0073] When constructing a model recognition model, it is necessary to back-transfer update parameters according to the classification function to obtain an optimized model recognition model.
[0074] Step 4: Based on the preset infusion bottle / infusion bag weight that matches the model and the identified infusion bottle / infusion bag model, determine the empty weight of the infusion bottle / infusion bag at the patient's location.
[0075] For each model of infusion bottle / infusion bag, an empty weight is preset. When applied, the infusion bottle / infusion bag model of the patient at the current position is obtained according to step 3, and the empty weight of the infusion bottle / infusion bag at the current patient's position is obtained.
[0076] Step 5: Based on the empty weight of the infusion bottle / infusion bag at each patient's location and the collected weight data of the infusion bottle / infusion bag at each patient's location, obtain the infusion progress of the current patient.
[0077] The infusion progress P of the current patient is:
[0078]
[0079] Among them, W in is the initial weight data of the infusion bottle / infusion bag at the patient’s location, W c The current weight data of the infusion bottle / infusion bag at the patient’s location, W e The empty weight of the infusion bottle / bag at the patient's location.
[0080] The initial weight data of the infusion bottle / infusion bag at the patient's location is the weight data of the infusion bottle / infusion bag collected when the current location changes from no patient to patient; that is, the weight data at the current location is collected. If it is lower than the set threshold, it indicates that there is no patient at the current location. The collection continues. When the collected weight data is greater than the set threshold for the first time, it means that there is a patient at the current location at the current collection moment. The weight data collected at this time is the initial weight data of the infusion bottle / infusion bag at the patient's location.
[0081] Step 6: Display the infusion progress of the current patient on the display terminal of the medical worker, and when the infusion progress of the current patient is greater than the set threshold, issue a reminder in a silent manner.
[0082] The display terminals include the medical workers’ PDAs and the central control screen in the infusion room.
[0083] When the reminder is issued in a silent manner, it includes but is not limited to vibrating the medical worker's PDA.
[0084] As an option, when the infusion progress reaches 98%, a reminder is issued. Of course, the reminder threshold of the infusion progress can also be adjusted according to expert experience.
[0085] The present invention automatically identifies the model of the infusion bottle / infusion bag through the model recognition model, and does not require medical workers to manually set the threshold or manually set the initial weight of the infusion bottle / infusion bag, thereby effectively reducing the workload of medical workers. At the same time, the infusion progress of the current patient can be automatically obtained according to the collected weight of the infusion bottle / infusion bag, which is convenient for medical staff to monitor the infusion progress and has a high degree of automation. In addition, the monitoring of the infusion by the present invention is in a silent state, which is conducive to creating a silent medical environment and improving the patient's medical experience.
[0086] In another embodiment, if Figure 4 As shown, a silent infusion monitoring system is provided, comprising:
[0087] Weighing module, used to collect weight data of infusion bottles / infusion bags at various locations in the infusion room;
[0088] An image acquisition module, used to acquire an infusion image of the patient's location;
[0089] A model recognition module is used to identify the model of the infusion bottle / infusion bag of the patient at the current location based on the infusion image of the patient's location and using the constructed model recognition model;
[0090] An empty weight determination module, for determining the empty weight of the infusion bottle / infusion bag at the patient's location based on a preset infusion bottle / infusion bag weight matching the model and the identified infusion bottle / infusion bag model;
[0091] An infusion progress determination module, used to obtain the infusion progress of the current patient based on the empty weight of the infusion bottle / infusion bag at each patient's location and the collected weight data of the infusion bottle / infusion bag at each patient's location;
[0092] The display module is used to display the infusion progress of the current patient on the display terminal of the medical worker, and to issue a reminder in a silent manner when the infusion progress of the current patient is greater than a set threshold.
[0093] For more specific processes of the above method, please refer to the corresponding contents disclosed in the aforementioned embodiments, which will not be repeated here.
[0094] In another embodiment, the present invention provides a computer device, including a processor and a memory; wherein the processor implements the steps of the above-mentioned silent infusion monitoring method when executing the computer program stored in the memory.
[0095] For more specific processes of the above method, please refer to the corresponding contents disclosed in the aforementioned embodiments, which will not be repeated here.
[0096] In another embodiment, the present invention provides a computer-readable storage medium for storing a computer program; when the computer program is executed by a processor, the steps of the above-mentioned silent infusion monitoring method are implemented.
[0097] For more specific processes of the above method, please refer to the corresponding contents disclosed in the aforementioned embodiments, which will not be repeated here.
[0098] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system, device, and storage medium disclosed in the embodiment, since they correspond to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.
[0099] Those skilled in the art can clearly understand that the technology in the embodiments of the present invention can be implemented by means of software plus a necessary general hardware platform. Based on this understanding, the technical solution in the embodiments of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a disk, an optical disk, etc., and includes a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention or some parts of the embodiments.
[0100] The above are only preferred embodiments of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should be regarded as the protection scope of the present invention.
Claims
1. A silent infusion monitoring method, characterized in that: include: Collect weight data of infusion bottles / infusion bags at various locations in the infusion room; Acquire an infusion image of the patient's location; Based on the infusion image of the patient's location, the constructed model recognition model is used to identify the model of the infusion bottle / infusion bag of the patient at the current location; Determine the empty weight of the infusion bottle / infusion bag at the patient's location based on the preset infusion bottle / infusion bag weight that matches the model and the identified infusion bottle / infusion bag model; Based on the empty weight of the infusion bottle / infusion bag at each patient's location and the collected weight data of the infusion bottle / infusion bag at each patient's location, the infusion progress of the current patient is obtained; The infusion progress of the current patient is displayed on the display terminal of the medical worker, and when the infusion progress of the current patient is greater than the set threshold, a reminder is issued in a silent manner.
2. The silent infusion monitoring method according to claim 1, characterized in that: The collecting of the infusion image of the patient's location includes: The current weight data of the collected infusion bottles / infusion bags at each location are compared with the set minimum threshold. If the current weight data of the infusion bottle / infusion bag at the current location is greater than the set minimum threshold, it indicates that there is a patient at the current location, and the infusion image is collected for the patient's location. Otherwise, there is no patient at the current location, and the infusion image is not collected for the current location.
3. The silent infusion monitoring method according to claim 1, characterized in that: The model recognition model is constructed using the Transformer algorithm; The infusion image based on the patient's location, using the constructed model recognition model to identify the model of the infusion bottle / infusion bag of the patient at the current location, includes: Step 3.1: The collected infusion image at the patient's location is sequentially divided into blocks and flattened to obtain an image patch vector; Step 3.2: Input all image patch vectors into the current Transformer Block stacking layer for feature extraction to obtain feature blocks containing feature information; Step 3.3: Perform Patch Merging on the feature map corresponding to the feature block, merge adjacent feature blocks, and input the merged feature block into the next Transformer Block stacking layer for feature extraction; Step 3.4: Repeat step 3.3 several times until all feature information is extracted; Step 3.5: Perform one-dimensional adaptive average pooling on all extracted feature information to obtain a global feature vector; and based on the global feature vector, classify through the Softmax classifier to achieve model recognition of the infusion bottle / infusion bag.
4. The silent infusion monitoring method according to claim 3, characterized in that: The step 3.1 comprises: Step 3.1.1: preprocessing the acquired infusion image at the patient's location; the preprocessing includes cutting, brightening and contrast enhancement; Step 3.1.2: Input the preprocessed infusion image into the Patch Partition module for partitioning; Step 3.1.3: For each block after the image is divided, flatten it in the channel direction to obtain the image patch vector.
5. The silent infusion monitoring method according to claim 3, characterized in that: In step 3.2, before all image patch vectors are input into the current Transformer Block stacking layer, the dimensions of the image patch vectors are adjusted through a linear Embedding layer; the Transformer Block stacking layer includes a window multi-head self-attention layer and a shifted window multi-head self-attention layer.
6. The silent infusion monitoring method according to claim 1, characterized in that: The infusion progress of the current patient is obtained based on the empty weight of the infusion bottle / infusion bag at each patient's location and the collected weight data of the infusion bottle / infusion bag at each patient's location. The infusion progress P of the current patient is: Among them, W in is the initial weight data of the infusion bottle / infusion bag at the patient’s location, W c The current weight data of the infusion bottle / infusion bag at the patient’s location, W e The empty weight of the infusion bottle / bag at the patient's location; The infusion progress of the current patient is displayed on the display terminal of the medical worker, and when the infusion progress of the current patient is greater than the set threshold, a reminder is issued in a silent manner. The display terminal includes a PDA of a medical worker and a central control screen of an infusion room.
7. A silent infusion monitoring system, characterized in that: include: Weighing module, used to collect weight data of infusion bottles / infusion bags at various locations in the infusion room; An image acquisition module, used to acquire an infusion image of the patient's location; A model recognition module is used to identify the model of the infusion bottle / infusion bag of the patient at the current location based on the infusion image of the patient's location and using the constructed model recognition model; An empty weight determination module, for determining the empty weight of the infusion bottle / infusion bag at the patient's location based on a preset infusion bottle / infusion bag weight matching the model and the identified infusion bottle / infusion bag model; An infusion progress determination module, used to obtain the infusion progress of the current patient based on the empty weight of the infusion bottle / infusion bag at each patient's location and the collected weight data of the infusion bottle / infusion bag at each patient's location; The display module is used to display the infusion progress of the current patient on the display terminal of the medical worker, and to issue a reminder in a silent manner when the infusion progress of the current patient is greater than a set threshold.
8. A computer device, characterized in that: It comprises a processor and a memory; wherein, when the processor executes the computer program stored in the memory, the steps of the silent infusion monitoring method described in any one of claims 1 to 6 are implemented.
9. A computer-readable storage medium, characterized in that: Used to store computer programs; when the computer programs are executed by the processor, the steps of the silent infusion monitoring method described in any one of claims 1-6 are implemented.
