Control method and device for a drug delivery device

Abstract

The application discloses a control method and device of a drug delivery device, wherein a current entity photograph of a target entity, first electromyography information, muscle sensor information, a target MRI tomogram and first electroencephalogram information are acquired; a target abnormality category is determined based on the current entity photograph, the first electromyography information, the muscle sensor information, the target MRI tomogram and the first electroencephalogram information; a target drug delivery category of the target entity is determined based on the target abnormality category of the target entity; drug delivery strategy information of the target entity is determined based on the target drug delivery category and entity information of the target entity, the drug delivery strategy information including a drug delivery time and a drug delivery amount; and a drug of the target drug delivery category is output by the drug delivery device based on the drug delivery strategy information. The application can improve drug delivery efficiency.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and specifically to a control method and apparatus for a drug delivery device. Background Technology

[0002] In the current technology, when patients use medication, the medication needs to be adjusted in a timely manner according to the patient's condition. However, at present, staff can only manually determine whether the medication needs to be changed, what kind of medication to change it to, and obtain the medication based on the patient's condition and examination reports. This manual method results in low medication administration efficiency. Summary of the Invention

[0003] This application provides a control method and apparatus for a drug delivery device, which can improve drug delivery efficiency.

[0004] In a first aspect, the control method for the drug delivery device provided in this application includes:

[0005] Acquire the current entity image, first electromyography information, muscle sensor information, target MRI tomography image, and first electroencephalography information of the target entity;

[0006] The target abnormality category is determined based on the current physical image, first electromyography information, muscle sensor information, target MRI tomography image, and first electroencephalography information.

[0007] Determine the target drug administration category for the target entity based on the target anomaly category of the target entity;

[0008] Based on the target drug administration category and the entity information of the target entity, the drug administration strategy information of the target entity is determined, and the drug administration strategy information includes the drug administration time and dosage;

[0009] Based on the drug delivery strategy information, the drug delivery device is controlled to output the drug of the target drug delivery category.

[0010] Optionally, the drug delivery device is disposed on a cuff device, the cuff device including multiple sensors for acquiring muscle sensor information, the drug delivery device including a universal tubing, multiple drug reservoirs, multiple power chambers, and a needle-free injection head, the multiple drug reservoirs and multiple power chambers corresponding one-to-one, the multiple drug reservoirs for storing different types of drugs, and the step of controlling the drug delivery device to output the drug of the target drug category based on the drug delivery strategy information includes:

[0011] Based on the target drug administration category of the target entity, the power output of the power chamber corresponding to the target drug administration category of the target entity is controlled to output the drug in the corresponding drug storage chamber through the needleless injection head.

[0012] Optionally, determining the target abnormality category based on the current entity image, first electromyography information, muscle sensor information, target MRI tomography image, and first electroencephalography information includes:

[0013] The capture time of the image of the current entity is taken as the current time;

[0014] When the acquisition end time of the first electromyography information is earlier than the current time, and the time difference between the acquisition end time of the first electromyography information and the current time is greater than the first preset duration, it is determined whether the number of channels in the first electromyography information exceeds n.

[0015] If the number of channels in the first electromyography information exceeds n, then the information of n preset channels in the first electromyography information is determined as the second electromyography information;

[0016] The muscle sensor information and the second electromyography (EMG) information are input into the target EMG generation model to obtain the third EMG information.

[0017] The target abnormality category is determined based on the current physical image, third electromyography information, muscle sensor information, target MRI tomography image, and first electroencephalography information.

[0018] Optionally, determining the target abnormality category based on the current entity image, third electromyography information, muscle sensor information, target MRI tomography image, and first electroencephalography information includes:

[0019] When the acquisition end time of the first EEG information is earlier than the current time, and the time difference between the acquisition end time of the first EEG information and the current time is greater than the second preset duration, a first classification model trained based on the first training set is obtained. The first classification model includes a first feature extraction module, a second feature extraction module, and a first classification module. The first training set includes multiple first samples and corresponding labeled abnormal categories. The first samples include entity image samples and corresponding EEG samples.

[0020] The current entity image is input into the first feature extraction module of the first classification model to obtain the first extracted features;

[0021] The first EEG information is input into the second feature extraction module of the first classification model to obtain the second extracted features;

[0022] The first and second extracted features are input into the target EEG generation model to obtain the second EEG information;

[0023] The target abnormality category is determined based on the current physical image, third electromyography information, muscle sensor information, target MRI tomography image, and second electroencephalography information.

[0024] Optionally, determining the target abnormality category based on the current entity image, third electromyography information, muscle sensor information, target MRI tomography image, and second electroencephalography information includes:

[0025] Obtain a second classification model trained on a second training set, wherein the second classification model includes a third feature extraction module and a second classification module, and the second training set includes multiple MRI tomographic image samples and corresponding labeled abnormal categories;

[0026] Input the target MRI tomographic image into the target lesion segmentation model to obtain multiple lesion segmentation regions on the target MRI tomographic image;

[0027] Each lesion segmentation region is input into the third feature extraction module of the second classification model to obtain multiple segmentation region features corresponding to multiple lesion segmentation regions;

[0028] By concatenating the features of multiple segmented regions, we obtain the feature information of the concatenated region;

[0029] The target abnormality category is determined based on the current physical image, third electromyography information, muscle sensor information, spliced ​​region feature information, and second electroencephalography information.

[0030] Optionally, the control method for the drug delivery device includes:

[0031] Multiple electromyography (EMG) information segments and their corresponding category labels are obtained. Each EMG information segment includes EMG information from N channels, where N is greater than n.

[0032] The electromyography (EMG) information fragment is cropped into multiple EMG information window segments based on a sliding window of preset length, resulting in multiple EMG information window segments and corresponding category labels;

[0033] The third classification model is trained by using multiple electromyography information windows and corresponding category labels of each channel as the third training set, and the first prediction accuracy of the third classification model is determined to obtain N first prediction accuracies for N channels.

[0034] The N1 channels corresponding to the top N1 first prediction accuracies, sorted from largest to smallest, are determined as N1 candidate channels, where n <N1<N;

[0035] Select n channels from N1 candidate channels as n preset channels.

[0036] Optionally, selecting n channels from N1 candidate channels as n preset channels includes:

[0037] For each of the multiple different entities, the electromyography information window segments of every n channels in the N1 candidate channels are combined into a training sample, resulting in multiple different training samples forming a fourth training set.

[0038] The fourth classification model is trained based on the fourth training set;

[0039] The electromyography information window segments of N1 candidate channels of the target entity are combined into a second sample for every n channels, resulting in multiple second samples and corresponding sample labels;

[0040] Each set of n channels is combined into a single channel, resulting in multiple n-channel combinations. Each n-channel combination contains a sample set, and each sample set includes multiple second samples belonging to the same n-channel combination.

[0041] Based on the fourth classification model, the sample set of each n-channel combination is predicted, and the second prediction accuracy of each n-channel combination is obtained.

[0042] The n channels in the n-channel combination with the highest prediction accuracy are selected as the n preset channels.

[0043] Secondly, the control device for the drug delivery device provided in this application includes:

[0044] The acquisition module is used to acquire the current entity image captured by the target entity, the first electromyography information, the muscle sensor information, the target MRI tomography image, and the first electroencephalogram information.

[0045] The first determining module is used to determine the target abnormality category based on the current entity image, the first electromyography information, the muscle sensor information, the target MRI tomography image, and the first electroencephalography information.

[0046] The second determination module is used to determine the target drug administration category of the target entity based on the target anomaly category of the target entity;

[0047] The third determining module is used to determine the dosing strategy information of the target entity based on the target drug category and the entity information of the target entity, wherein the dosing strategy information includes the dosing time and dosage.

[0048] The control module is used to control the drug delivery device to output the drug of the target drug delivery category based on the drug delivery strategy information.

[0049] Thirdly, the electronic device provided in this application includes a memory and a processor. The memory stores a computer program, and the processor is used to run the computer program in the memory to implement the steps in the control method of the drug delivery device provided in this application.

[0050] Fourthly, the computer-readable storage medium provided in this application stores a plurality of instructions that are adapted for loading by a processor to implement the steps in the control method of the drug delivery device provided in this application.

[0051] Fifthly, the computer program product provided in this application includes a computer program or instructions that, when executed by a processor, implement the steps in the control method of the drug delivery device provided in this application.

[0052] In this application, compared to related technologies, the following methods are employed: acquiring a current entity image, first electromyography (EMG) information, muscle sensor information, target MRI tomography image, and first electroencephalography (EEG) information of the target entity; determining the target abnormality category based on the current entity image, first EMG information, muscle sensor information, target MRI tomography image, and first EEG information; determining the target drug delivery category based on the target abnormality category of the target entity; determining the drug delivery strategy information of the target entity based on the target drug delivery category and the entity information of the target entity, the drug delivery strategy information including drug delivery time and dosage; and controlling the drug delivery device to output the drug of the target drug delivery category based on the drug delivery strategy information. This application can improve drug delivery efficiency. Attached Figure Description

[0053] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0054] Figure 1 This is a schematic diagram of a control system for the drug delivery device provided in an embodiment of this application;

[0055] Figure 2 This is a schematic diagram of the sleeve device in the control system of the drug delivery device provided in the embodiments of this application;

[0056] Figure 3 This is a schematic diagram of the structure of the drug delivery device in the control system of the drug delivery device provided in the embodiments of this application;

[0057] Figure 4 This is a schematic flowchart of an embodiment of the control method for the drug delivery device provided in this application.

[0058] Figure 5This is a schematic diagram of the structure of the first classification model in one embodiment of the control method for the drug delivery device provided in this application;

[0059] Figure 6 This is a flowchart illustrating another embodiment of the control method for the drug delivery device provided in this application.

[0060] Figure 7 This is a schematic diagram of the structure of the control device of the drug delivery device provided in the embodiments of this application;

[0061] Figure 8 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0062] It should be noted that the principles of this application are illustrated by example in a suitable computing environment. The following description is based on the specific embodiments of this application that are illustrated, and should not be regarded as limiting other specific embodiments not detailed herein.

[0063] In the following description of this application, "some embodiments" are referred to, which describe a subset of all possible embodiments. However, it is understood that "some embodiments" may be the same subset or different subset of all possible embodiments, and may be combined with each other without conflict.

[0064] In the following description of this application, the terms "first, second, third" are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first, second, third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0065] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0066] To improve the control effect of a drug delivery device, embodiments of this application provide a control method for a drug delivery device, a control device for a drug delivery device, an electronic device, a computer-readable storage medium, and a computer program product. The control method for the drug delivery device can be executed by the control device for the drug delivery device, or by an electronic device integrating the control device for the drug delivery device.

[0067] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0068] Please refer to Figure 1 This application also provides a control system for a drug delivery device, such as Figure 1 As shown, the drug delivery device includes a control system electronic device 100 and a sleeve device. The electronic device 100 integrates the control device of the drug delivery device provided in this application. The electronic device 100 and the sleeve device communicate to send and receive commands.

[0069] like Figure 2 and Figure 3 As shown in the embodiment of this application, the drug delivery device is disposed on a sleeve device. The sleeve device includes multiple sensors for acquiring muscle sensor information. The drug delivery device includes a universal tubing, multiple drug storage chambers, multiple power chambers, and a needle-free injection head. The multiple drug storage chambers and multiple power chambers correspond one-to-one, and the multiple drug storage chambers are used to store different types of drugs. The needle-free injection head is also provided with an injection port.

[0070] The sleeve device has an elastic and stretchable function and can be worn on the arms, legs, upper limbs, torso, etc. This application uses wearing the sleeve on the upper arm as an example for illustration. The sleeve device has a grid-like shape, and there are multiple sensors in the grid. These sensors can be pulse pressure sensors, body surface temperature sensors, muscle tension sensors, and force-exerting muscle position sensors. The multiple sensors acquire p rows of time-series data. Here, we take these 5 rows of data as an example. If more types of data need to be measured, the number of rows can be more than 5.

[0071] Figure 3 Taking a three-compartment administration device as an example, the administration device consists of a power chamber, a medication compartment, a universal tubing, a needle-free injection head, and an injection port. The power chamber serves as the driving mechanism, powered by a compression spring or high-pressure gas. The power unit is a disposable consumable; the compression spring or compressed gas needs to be refilled after each use. The medication compartment must be able to hold at least five times the adult's required dosage of this type of medication. The medication is a consumable and needs to be refilled after each use to prepare for the next use. If the medication has not been used for a long time, the old medication should be periodically removed and replaced with new medication. The universal tubing is a disposable consumable and needs to be replaced after each use. It is arranged horizontally within the sleeve, connecting multiple medication tubings to the needle-free injection head. Regardless of which medication tubing is activated, the medication uses this universal tubing. The needle-free injection head is a disposable consumable and needs to be replaced after each use. It is arranged vertically and fits snugly against the skin surface. The medication is dispensed through the injection port.

[0072] Among them, electronic device 100 can be any device equipped with a processor and having processing capabilities, such as mobile electronic devices with processors such as smartphones, tablets, PDAs, laptops, and smart speakers, or fixed electronic devices with processors such as desktop computers, televisions, servers, and industrial equipment.

[0073] In addition, such as Figure 1 As shown, the control system of the drug delivery device may also include a memory for storing raw data, intermediate data, and result data.

[0074] In this embodiment of the application, the storage device can be a cloud storage device. Cloud storage is a new concept that is extended and developed from the concept of cloud computing. A distributed cloud storage system (hereinafter referred to as a storage system) refers to a storage system that uses cluster applications, grid technology and distributed storage file system functions to bring together a large number of storage devices of various types in the network (storage devices are also called storage nodes) through application software or application interfaces to work together to provide data storage and business access functions to the outside world.

[0075] Currently, the storage method of storage systems is as follows: Logical volumes are created. During the creation of a logical volume, physical storage space is allocated to each logical volume. This physical storage space may consist of a single storage device or the disks of several storage devices. Clients store data on a logical volume, which means storing the data on the file system. The file system divides the data into many parts, each part being an object. Each object contains not only the data but also additional information such as a data identifier (ID, ID entity). The file system writes each object to the physical storage space of that logical volume and records the storage location information of each object. Therefore, when a client requests access to data, the file system can allow the client to access the data based on the storage location information of each object.

[0076] The process by which a storage system allocates physical storage space to a logical volume is as follows: the physical storage space is pre-divided into strips according to the capacity estimate of the objects stored in the logical volume (this estimate often has a large margin relative to the actual capacity of the objects to be stored) and the grouping of Redundant Array of Independent Disks (RAID). A logical volume can be understood as a strip, thus allocating physical storage space to the logical volume.

[0077] It should be noted that, Figure 1The schematic diagram of the control system of the drug delivery device shown is merely an example. The control system and scenario of the drug delivery device described in the embodiments of this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided in the embodiments of this application. As those skilled in the art will know, with the evolution of the control system of the drug delivery device and the emergence of new business scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0078] The following sections provide detailed descriptions of each example. It should be noted that the sequence numbers of the following embodiments are not intended to limit the preferred order of the embodiments.

[0079] Please refer to Figure 4 , Figure 4 This is a schematic flowchart of one embodiment of the control method for the drug delivery device provided in this application, as shown below. Figure 4 As shown, the flow of the control method for the drug delivery device provided in this application is as follows:

[0080] 201. Acquire the current entity image, first electromyography information, muscle sensor information, target MRI tomography image, and first electroencephalography information of the target entity.

[0081] In this embodiment, entity-captured video footage is obtained from a camera. This video footage includes the limb movements of the target entity. During the day, the footage is captured using a visible light camera (r / g / b three-channel), and at night, it is captured using an infrared camera (single-channel). The entity-captured video is processed frame by frame to obtain multiple current entity-captured images of the target entity. The current entity-captured image obtained at the current time is then acquired. The current time is the time when the current entity-captured image was captured.

[0082] In this embodiment of the application, the first electromyography information can be the electromyography signal of the target entity currently being collected, or the first electromyography information can be the electromyography signal of the target entity collected in the past.

[0083] In this embodiment of the application, the first EEG information can be the EEG signal of the target entity currently being collected, or the first EEG information can be the EEG signal of the target entity collected in the past.

[0084] In this embodiment, the drug delivery device is disposed on a sleeve device, which includes multiple sensors for acquiring muscle sensor information. The drug delivery device includes a universal tubing, multiple drug storage chambers, multiple power chambers, and a needle-free injection head. The multiple drug storage chambers and multiple power chambers correspond one-to-one, and the multiple drug storage chambers are used to store different types of drugs.

[0085] The sleeve device has an elastic and stretchable function and can be worn on the arms, legs, upper limbs, torso, etc. This application uses wearing the sleeve on the upper arm as an example for illustration. The sleeve device has a grid-like shape, and there are multiple sensors in the grid. These sensors can be pulse pressure sensors, body surface temperature sensors, muscle tension sensors, and force-exerting muscle position sensors. The multiple sensors acquire p rows of time-series data. Here, we take these 5 rows of data as an example. If more types of data need to be measured, the number of rows can be more than 5.

[0086] Specifically, pulse pressure data, body surface temperature data, muscle tension data, and muscle position data are acquired and then constructed into five time-series signals: pulse pressure, temperature, tension, x, and y, which serve as muscle sensor information.

[0087] In this embodiment of the application, the target MRI tomographic image is the target MRI tomographic image obtained from the most recent MRI tomographic scan of the target entity.

[0088] 202. Determine the target abnormality category based on the current entity image, first electromyography information, muscle sensor information, target MRI tomography image, and first electroencephalography information.

[0089] In this embodiment of the application, the target abnormality category can be primary myoclonus, subcortical non-segmental myoclonus, myoclonic dystonia, reticular reflex myoclonus, or spinal myoclonus.

[0090] In one specific embodiment, an anomaly classification model is trained by inputting the current entity's captured image, first electromyography (EMG) information, muscle sensor information, target MRI tomography image, and first electroencephalography (EEG) information into the anomaly classification model to obtain the target anomaly category.

[0091] In another specific embodiment, the target abnormality category is determined based on the current entity image, first electromyography (EMG) information, muscle sensor information, target MRI tomography image, and first electroencephalography (EEG) information, including:

[0092] (1) Use the time when the current entity's image is captured as the current time.

[0093] (2) When the acquisition end time of the first electromyography information is earlier than the current time, and the time difference between the acquisition end time of the first electromyography information and the current time is greater than the first preset duration, determine whether the number of channels in the first electromyography information exceeds n.

[0094] The first preset duration can be 1 week, 2 weeks, or other values.

[0095] When the acquisition end time of the first electromyography (EMG) data is earlier than the current time, and the time difference between the acquisition end time of the first EMG data and the current time is greater than the first preset duration, it indicates that the acquisition time of the first EMG data is too long and may not be accurate, and more accurate EMG data needs to be generated.

[0096] When the acquisition end time of the first electromyography (EMG) information is earlier than the current time and the time difference between the acquisition end time of the first EMG information and the current time is not greater than the first preset duration, it indicates that the acquisition time of the first EMG information is short and it can be used. Then, the current entity image, the first EMG information, the muscle sensor information, the target MRI tomographic image, and the first EEG information are input into the abnormal classification model to obtain the target abnormal category.

[0097] (3) If the number of channels in the first electromyography information exceeds n, then the information of n preset channels in the first electromyography information is determined as the second electromyography information.

[0098] n preset channels can be pre-set.

[0099] In this embodiment of the application, the control method for the drug delivery device includes:

[0100] Step 1-1: Obtain multiple electromyography (EMG) information fragments and their corresponding category labels. Each EMG information fragment includes EMG information from N channels.

[0101] Specifically, electromyography (EMG) signals from specific patients are tracked and collected. The EMG signals have N channels, typically N≥8. Based on the patient's myoclonus / non-myoclonus status, the collected EMG signals are cropped into multiple large segments, resulting in multiple EMG information fragments. If the current segment occurs during myoclonus, the corresponding EMG information fragment is categorized as myoclonus; if the current segment occurs during non-myoclonus, the corresponding EMG information fragment is categorized as non-myoclonus.

[0102] Furthermore, the category labels can be primary myoclonus, subcortical non-segmental myoclonus, myoclonic dystonia, reticular reflex myoclonus, and spinal myoclonus.

[0103] Steps 1-2: Based on a sliding window of preset length, the electromyography (EMG) information fragment is cropped into multiple EMG information window segments, resulting in multiple EMG information window segments and corresponding category labels.

[0104] The preset length of the sliding window can be 500 / 600 / 1000 points, etc., and this application does not impose a specific limitation. For example, each electromyography (EMG) information segment is divided into 10 EMG information window segments, and 10 EMG information segments are further divided into 100 EMG information window segments.

[0105] Steps 1-3: Train the third classification model by using multiple electromyography information windows and corresponding category labels of each channel as the third training set, and determine the first prediction accuracy of the third classification model to obtain N first prediction accuracies for N channels.

[0106] In this embodiment, multiple electromyography information windows and corresponding sample labels for each channel are used as a third training set, resulting in N third training sets. A third classification model is trained using these N third training sets, yielding the first prediction accuracy of each of the N third classification models, and thus the N first prediction accuracies for each of the N channels. The first prediction accuracy can be considered the accuracy metric of the third classification model on the third training sets. Accuracy is a metric used to evaluate classification models.

[0107] Steps 1-4: The top N1 channels corresponding to the N first prediction accuracies, sorted from largest to smallest, are determined as N1 candidate channels, where n <N1<N。

[0108] For example, N is 8, N1 is 7, and n is 5; these can be set according to the specific situation.

[0109] Steps 1-5: Select n channels from N1 candidate channels as n preset channels.

[0110] In a specific embodiment, n channels are randomly selected from N1 candidate channels as n preset channels.

[0111] In another specific embodiment, selecting n channels from N1 candidate channels as n preset channels includes:

[0112] Step 2-1: For each of the multiple different entities, combine the electromyography information windows of every n channels in the N1 candidate channels into a training sample, and obtain multiple different training samples to form the fourth training set.

[0113] In this embodiment of the application, for one of multiple different entities, an electromyography information window segment of n channels from N1 candidate channels is selected as a training sample, i.e. The fourth training set is composed of all randomly paired training samples. For example, if there are k entities and m window segments in each channel, then the fourth training set has... Training samples.

[0114] Step 2-2: Train the fourth classification model based on the fourth training set.

[0115] Steps 2-3: Combine the electromyography information window segments of every n channels in the N1 candidate channels of the target entity into a second sample, and obtain multiple second samples and corresponding sample labels.

[0116] For the target entity, select n channels of electromyography information window segments from N1 candidate channels as a second sample, i.e. There are m second samples. For example, if each channel has m window segments, then there are... A second sample.

[0117] Steps 2-4: Combine every n channels as a channel combination to obtain multiple n-channel combinations. Obtain a sample set for each n-channel combination. Each sample set includes multiple second samples belonging to the same n-channel combination.

[0118] Specifically, the sample set for each n-channel combination contains m second samples.

[0119] Steps 2-5: Based on the fourth classification model, predict the sample set of each n-channel combination to obtain the second prediction accuracy of each n-channel combination.

[0120] Specifically, for a sample set of one of the n-channel combinations, there are 10 second samples. If the fourth classification model correctly predicts 9 of them, then the second prediction accuracy of the n-channel combination is 90%.

[0121] Steps 2-6: Select the n channels from the n-channel combination with the highest prediction accuracy as the n preset channels.

[0122] (4) Input the muscle sensor information and the second electromyography information into the target electromyography generation model to obtain the third electromyography information.

[0123] Of course, in other embodiments, only muscle sensor information can be input into the target electromyography (EMG) generation model to obtain third EMG information.

[0124] In this embodiment of the application, the target electromyography generation model can be a trained generative adversarial network (CGAN).

[0125] Specifically, after aligning the muscle sensor information with the second electromyography (EMG) information based on the timestamp, the information is input into the target EMG generation model to obtain the third EMG information with n channels.

[0126] (5) Determine the target abnormality category based on the current entity image, third electromyography information, muscle sensor information, target MRI tomography image and first electroencephalogram information.

[0127] In one specific embodiment, an anomaly classification model is trained by inputting the current entity image, third electromyography information, muscle sensor information, target MRI tomography image, and first electroencephalography information into the anomaly classification model to obtain the target anomaly category.

[0128] In another specific embodiment, the target abnormality category is determined based on the current entity image, third electromyography information, muscle sensor information, target MRI tomography image, and first electroencephalography information, including:

[0129] Step 3-1: When the acquisition end time of the first EEG information is earlier than the current time, and the time difference between the acquisition end time of the first EEG information and the current time is greater than the second preset duration, obtain the first classification model trained based on the first training set. The first classification model includes a first feature extraction module, a second feature extraction module, and a first classification module. The first training set includes multiple first samples and corresponding labeled abnormal categories. The first samples include entity image samples and corresponding EEG samples.

[0130] like Figure 5 As shown, the first classification model includes a first feature extraction module, a second feature extraction module, and a first classification module. The first feature extraction module receives a video frame format image of the current entity through input 1 and outputs the first extracted feature, i.e., feature 1. The first feature extraction module includes multiple convolutional layers and feature deformation layers. The feature deformation layer performs a reshape operation. The first input parameter of input 1 is input_shape = (W1, H1, D1), where W1 is the image width of the current entity image, H1 is the image height of the current entity image, and D1 is the image depth of the current entity image. Specifically, D1 = 3 for visible light and D1 = 1 for infrared images. The second feature extraction module receives first EEG information in EEG window fragment format through input 2 and outputs the second extracted feature, i.e., feature 2. The second feature extraction module includes multiple convolutional layers and feature deformation layers. The feature deformation layer is used to perform reshape operations. The second input parameter of input 2 is input_shape = (W2, H2, 1), where W2 is the number of points contained in the first EEG information, and H2 is the number of rows in the first EEG information. The first classification module is used for feature concatenation and classification.

[0131] The second preset duration can be 1 week, 2 weeks, or other values.

[0132] When the collection time of the first EEG information ends earlier than the current time, and the time difference between the collection time of the first EEG information and the current time is greater than the second preset duration, it indicates that the collection time of the first EEG information is too long and may not be accurate, and more accurate EEG information needs to be generated.

[0133] When the acquisition end time of the first EEG information is earlier than the current time and the time difference between the acquisition end time of the first EEG information and the current time is not greater than the first preset duration, it indicates that the acquisition time of the first EEG information is short and it can be used. Then, the current entity image, the third electromyography information, the muscle sensor information, the target MRI tomography image and the first EEG information are input into the abnormal classification model to obtain the target abnormal category.

[0134] The first training set includes multiple first samples and corresponding labeled anomaly categories. The first samples include entity image samples and corresponding electroencephalogram (EEG) samples. The first training set can be manually collected and labeled.

[0135] In this embodiment, the entity image sample in the first sample is input into the first feature extraction module of the first classification model, and the electroencephalogram sample in the first sample is input into the second feature extraction module of the first classification model to obtain the predicted abnormal category output by the first classification module of the first classification model; the loss value is calculated based on the predicted abnormal category and the labeled abnormal category of the first sample; the first classification model is iteratively trained based on the loss value to obtain the first classification model trained on the first training set.

[0136] Specifically, the loss value can be calculated based on the predicted anomaly category and the labeled anomaly category of the first sample using the cross-entropy loss function. The first classification model can then be iteratively trained based on the loss value until the loss value is less than the preset loss value, thus obtaining the first classification model trained on the first training set.

[0137] Step 3-2: Input the current entity image into the first feature extraction module of the first classification model to obtain the first extracted features.

[0138] Step 3-3: Input the first EEG information into the second feature extraction module of the first classification model to obtain the second extracted features.

[0139] Steps 3-4: Input the first and second extracted features into the target EEG generation model to obtain the second EEG information.

[0140] Steps 3-5: Determine the target abnormality category based on the current entity image, third electromyography information, muscle sensor information, target MRI tomography image, and second electroencephalography information.

[0141] In one specific embodiment, an anomaly classification model is trained by inputting the current entity's captured image, third electromyography (EMG) information, muscle sensor information, target MRI tomography image, and second electroencephalography (EEG) information into the anomaly classification model to obtain the target anomaly category.

[0142] In another specific embodiment, the target abnormality category is determined based on the current entity image, third electromyography (EMG) information, muscle sensor information, target MRI tomography image, and second electroencephalography (EEG) information, including:

[0143] Step 4-1: Obtain the second classification model trained on the second training set. The second classification model includes a third feature extraction module and a second classification module. The second training set includes multiple MRI tomographic image samples and corresponding labeled abnormal categories.

[0144] The second classification model can be ResNet50 or VGG16.

[0145] Step 4-2: Input the target MRI tomographic image into the target lesion segmentation model to obtain multiple lesion segmentation regions on the target MRI tomographic image.

[0146] The target lesion segmentation model is a pre-trained model, such as Unet / Unet++.

[0147] Step 4-3: Input each lesion segmentation region into the third feature extraction module of the second classification model to obtain multiple segmentation region features corresponding to multiple lesion segmentation regions.

[0148] Step 4-4: Concatenate the features of multiple segmented regions to obtain the feature information of the concatenated region.

[0149] Steps 4-5: Determine the target abnormality category based on the current entity image, third electromyography information, muscle sensor information, splicing region feature information, and second electroencephalography information.

[0150] In this embodiment, the current entity image, stitched region feature information, and second EEG information are reshaped to generate a two-dimensional matrix deep3. The muscle sensor information deep1, the third EMG information deep2, and the two-dimensional matrix deep3 are then combined to form a three-channel image, which is input into a convolutional neural network for classification to obtain the target abnormality category. The classification labels of the convolutional neural network are primary myoclonus, subcortical non-segmental myoclonus, myoclonic dystonia, reticular reflex myoclonus, and spinal myoclonus.

[0151] 203. Determine the target drug administration category of the target entity based on the target anomaly category of the target entity.

[0152] In this embodiment, a pre-defined correspondence is established between drug administration categories and target abnormality categories. Specifically, the drug administration category is clonazepam, and the target abnormality categories are: primary myoclonus, subcortical non-segmental myoclonus, myoclonic dystonia, reticular reflex myoclonus, and spinal myoclonus, etc.; the drug administration category is immunoglobulin, and the target abnormality category is myoclonus caused by immune inflammation; the drug administration category is botulinum toxin, and the target abnormality categories are palatal myoclonus and segmental myoclonus; the drug administration category is serotonin, and the target abnormality category is hypoxic myoclonus. The specific drug required for each type of myoclonus must meet clinical requirements.

[0153] 204. Determine the dosing strategy information for the target entity based on the target dosing category and the entity information of the target entity. The dosing strategy information includes the dosing time and dosage.

[0154] In this embodiment, the dosage recommendation model is trained by inputting the entity information of the target entity and the target drug administration category into the model to obtain the dosage and administration time. Specifically, the entity information of the target entity includes the patient's gender (gen), age (age), weight (g), and medical history (T). A training set is formed by acquiring the entity information of multiple entities and their corresponding category labels (scores) to train the dosage recommendation model. The dosage recommendation model can be a decision tree, random forest, or backpropagation neural network, etc.

[0155] Load the dosage recommendation model to obtain the dosage l for each administration and generate dosing strategy information.

[0156] For example, the dosing strategy information includes: if the patient's myoclonus lasts for less than 5 minutes, no medication is given; if the patient's myoclonus lasts for more than 5 minutes, medication is given, with a total dosage of 1 liter per dose, injected slowly over approximately 30 seconds. After the first dose, if the condition is still not controlled, if the confidence level of the dosage recommendation model output is ≤τ1, it indicates that the condition is within a controllable range, and the original dose is repeated once every 20 minutes; if the confidence level of the dosage recommendation model output is ∈(τ1,τ2], it indicates that the condition is relatively serious, and the original dose is repeated twice every 20 minutes; if the confidence level of the dosage recommendation model output is >τ2, it is recommended to immediately send the patient to the hospital.

[0157] 205. Control the drug delivery device to output the target drug category based on drug delivery strategy information.

[0158] In this application, the power output of the power chamber corresponding to the target drug delivery category of the target entity is controlled based on the target drug delivery category of the target entity, and the drug in the corresponding drug storage chamber is output through the needleless injection head.

[0159] like Figure 6As shown in the embodiments of this application, the control method of the drug delivery device includes: acquiring electromyography (EMG), collecting muscle signals from the EMG cuff, and establishing a feature mapping 1 between the EMG and the muscle signals, for example, feature mapping 1 is a target EMG generation model. Acquiring electroencephalography (EEG), collecting limb movement features obtained from video monitoring, and establishing a feature mapping 2 between the EEG and the limb movement features, for example, feature mapping 2 is a target EEG generation model.

[0160] During the usage phase, muscle signals from the electromyography (EMG) cuff are collected, and a mapped muscle point map is generated using a target EMG generation model. Limb movement features obtained from video monitoring are also collected, and mapped EEG features are generated using a target EEG generation model. Image features are also acquired from head MRI. The mapped muscle point map, muscle signals, mapped EEG features, and image features acquired from head MRI are input into a PC. The PC then determines the type of myoclonus, issues commands, and controls the drug delivery device on the EMG cuff to output the drug based on the drug type and drug injection strategy.

[0161] To facilitate better implementation of the control method for the drug delivery device provided in the embodiments of this application, the embodiments of this application also provide a control device for the drug delivery device based on the above-described control method. The meanings of the terms used are the same as in the control method for the drug delivery device described above, and for specific implementation details, please refer to the descriptions in the above method embodiments.

[0162] Please refer to Figure 7 , Figure 7 This is a schematic diagram of the structure of the control device for the drug delivery device provided in the embodiments of this application. The control device for the drug delivery device may include:

[0163] The acquisition module 701 is used to acquire the current entity image captured by the target entity, the first electromyography information, the muscle sensor information, the target MRI tomography image, and the first electroencephalogram information.

[0164] The first determining module 702 is used to determine the target abnormality category based on the current entity image, the first electromyography information, the muscle sensor information, the target MRI tomography image, and the first electroencephalography information.

[0165] The second determining module 703 is used to determine the target drug administration category of the target entity based on the target anomaly category of the target entity;

[0166] The third determining module 704 is used to determine the dosing strategy information of the target entity based on the target drug category and the entity information of the target entity, wherein the dosing strategy information includes the dosing time and dosage.

[0167] The control module 705 is used to control the drug delivery device to output the drug of the target drug delivery category based on the drug delivery strategy information.

[0168] For details on the implementation of each of the above modules, please refer to the previous examples, which will not be repeated here.

[0169] This application also provides an electronic device, including a memory and a processor, wherein the processor executes steps in the control method of the drug delivery device provided in this embodiment by calling a computer program stored in the memory.

[0170] Please refer to Figure 8 , Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0171] The electronic device may include components such as a processor 101 with one or more processing cores, a memory 102 with one or more computer-readable storage media, a power supply 103, and an input unit 104. Those skilled in the art will understand that the electronic device structure shown in the figures does not constitute a limitation on the electronic device, and may include more or fewer components than shown, or combine certain components, or have different component arrangements. Wherein:

[0172] The processor 101 is the control center of the electronic device, connecting various parts of the device via various interfaces and lines. It executes software programs and / or modules stored in the memory 102, and calls data stored in the memory 102, to perform various functions and process data. Optionally, the processor 101 may include one or more processing cores; alternatively, the processor 101 may integrate an application processor and a modem processor, wherein the application processor mainly handles the operating system, user interface, and applications, while the modem processor mainly handles wireless communication. It is understood that the modem processor may not be integrated into the processor 101.

[0173] The memory 102 can be used to store software programs and modules. The processor 101 executes various functional applications and data processing by running the software programs and modules stored in the memory 102. The memory 102 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the electronic device, etc. In addition, the memory 102 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory 102 may also include a memory controller to provide the processor 101 with access to the memory 102.

[0174] The electronic device also includes a power supply 103 that supplies power to the various components. Optionally, the power supply 103 can be logically connected to the processor 101 through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system. The power supply 103 may also include one or more DC or AC power supplies, recharging systems, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components.

[0175] The electronic device may also include an input unit 104, which can be used to receive input digital or character information and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function control.

[0176] Although not shown, the electronic device may also include a display unit, an image acquisition component, etc., which will not be described in detail here. Specifically, in this embodiment, the processor 101 in the electronic device loads one or more executable codes corresponding to computer programs into the memory 102 according to the following instructions, and the processor 101 executes the steps in the control method of the drug delivery device provided in this application, such as:

[0177] The system acquires the current entity image, first electromyography (EMG) information, muscle sensor information, target MRI tomography image, and first electroencephalography (EEG) information of the target entity; determines the target abnormality category based on the current entity image, first EMG information, muscle sensor information, target MRI tomography image, and first EEG information; determines the target drug administration category of the target entity based on the target abnormality category; determines the drug administration strategy information of the target entity based on the target drug administration category and the entity information of the target entity, including drug administration time and dosage; and controls the drug administration device to output the drug of the target drug administration category based on the drug administration strategy information.

[0178] It should be noted that the electronic device provided in this application embodiment and the control method of the drug delivery device in the above embodiment belong to the same concept. The specific implementation process can be found in the above related embodiments, and will not be repeated here.

[0179] This application also provides a computer-readable storage medium storing a computer program thereon. When the computer program stored thereon is executed on the processor of the electronic device provided in the embodiments of this application, the processor of the electronic device performs the steps in the control method of the drug delivery device provided in this application. The storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM), etc.

[0180] This application also provides a computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform various optional implementations of the control method for the drug delivery device described above.

[0181] The control method and apparatus for a drug delivery device provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

[0182] It should be noted that when the above embodiments of this application are applied to specific products or technologies, and user-related data is involved, user permission or consent is required, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

Claims

1. A control device for a drug delivery device, characterized in that, include: The acquisition module is used to acquire the current entity image captured by the target entity, the first electromyography information, the muscle sensor information, the target MRI tomography image, and the first electroencephalogram information. The first determining module is used to determine the target abnormality category based on the current entity image, the first electromyography information, the muscle sensor information, the target MRI tomography image, and the first electroencephalography information. The capture time of the image of the current entity is taken as the current time; When the acquisition end time of the first electromyography information is earlier than the current time, and the time difference between the acquisition end time of the first electromyography information and the current time is greater than the first preset duration, it is determined whether the number of channels in the first electromyography information exceeds n. If the number of channels in the first electromyography (EMG) information exceeds n, the muscle sensor information is input into the target EMG generation model to obtain n-channel third EMG information. This involves acquiring multiple EMG information segments and corresponding category labels, where each EMG information segment includes EMG information from N channels, where N is greater than n. The EMG information segments are then cropped into multiple EMG information window segments based on a preset length sliding window, resulting in multiple EMG information window segments and corresponding category labels. The multiple EMG information window segments and corresponding category labels for each channel are used as a third training set to train the third classification model, and the first prediction accuracy of the third classification model is determined, resulting in N first prediction accuracies for N channels. The N first prediction accuracies are then sorted from largest to smallest, with the top-ranked ones being the most accurate. The first prediction accuracy corresponds to The channels were determined as follows There are 10 candidate channels, among which... ;from n channels are selected from the candidate channels as n preset channels; the target abnormality category is determined based on the current entity image, the third electromyography information, the muscle sensor information, the target MRI tomography image, and the first electroencephalogram information. When the acquisition end time of the first electromyography information is earlier than the current time and the time difference between the acquisition end time of the first electromyography information and the current time is not greater than the first preset duration, the current entity image, the first electromyography information, the muscle sensor information, the target MRI tomography image, and the first electroencephalogram information are input into the abnormality classification model to obtain the target abnormality category. The second determination module is used to determine the target drug administration category of the target entity based on the target anomaly category of the target entity; The third determining module is used to determine the dosing strategy information of the target entity based on the target drug category and the entity information of the target entity, wherein the dosing strategy information includes the dosing time and dosage. The control module is used to control the drug delivery device to output the drug of the target drug delivery category based on the drug delivery strategy information.

2. The control device for the drug delivery device according to claim 1, characterized in that, The drug delivery device is mounted on a sleeve device, which includes multiple sensors for acquiring muscle sensor information. The drug delivery device includes a universal tubing, multiple drug reservoirs, multiple power chambers, and a needle-free injection head. The multiple drug reservoirs and power chambers correspond one-to-one. The multiple drug reservoirs are used to store different types of drugs. Controlling the drug delivery device to output the target drug category based on the drug delivery strategy information includes: Based on the target drug administration category of the target entity, the power output of the power chamber corresponding to the target drug administration category of the target entity is controlled to output the drug in the corresponding drug storage chamber through the needleless injection head.

3. The control device for the drug delivery device according to claim 1, characterized in that, The process of determining the target abnormality category based on the current entity image, third electromyography information, muscle sensor information, target MRI tomography image, and first electroencephalography information includes: When the acquisition end time of the first EEG information is earlier than the current time, and the time difference between the acquisition end time of the first EEG information and the current time is greater than the second preset duration, a first classification model trained based on the first training set is obtained. The first classification model includes a first feature extraction module, a second feature extraction module, and a first classification module. The first training set includes multiple first samples and corresponding labeled abnormal categories. The first samples include entity image samples and corresponding EEG samples. The current entity image is input into the first feature extraction module of the first classification model to obtain the first extracted features; The first EEG information is input into the second feature extraction module of the first classification model to obtain the second extracted features; The first and second extracted features are input into the target EEG generation model to obtain the second EEG information; The target abnormality category is determined based on the current physical image, third electromyography information, muscle sensor information, target MRI tomography image, and second electroencephalography information.

4. The control device for the drug delivery apparatus according to claim 3, characterized in that, The process of determining the target abnormality category based on the current entity image, third electromyography (EMG) information, muscle sensor information, target MRI tomography image, and second electroencephalography (EEG) information includes: Obtain a second classification model trained on a second training set, wherein the second classification model includes a third feature extraction module and a second classification module, and the second training set includes multiple MRI tomographic image samples and corresponding labeled abnormal categories; Input the target MRI tomographic image into the target lesion segmentation model to obtain multiple lesion segmentation regions on the target MRI tomographic image; Each lesion segmentation region is input into the third feature extraction module of the second classification model to obtain multiple segmentation region features corresponding to multiple lesion segmentation regions; By concatenating the features of multiple segmented regions, we obtain the feature information of the concatenated region; The target abnormality category is determined based on the current physical image, third electromyography information, muscle sensor information, spliced ​​region feature information, and second electroencephalography information.

5. The control device for the drug delivery device according to claim 1, characterized in that, The from From the candidate channels, n channels are selected as n preset channels, including: For each of the multiple different entities, The electromyography information window segments of each of the n candidate channels are combined into a training sample, and multiple different training samples are combined to form a fourth training set. The fourth classification model is trained based on the fourth training set; Obtain the target entity The electromyography information window segments of each of the n candidate channels are combined into a second sample, resulting in multiple second samples and corresponding sample labels; Each set of n channels is combined into a single channel, resulting in multiple n-channel combinations. Each n-channel combination contains a sample set, and each sample set includes multiple second samples belonging to the same n-channel combination. Based on the fourth classification model, the sample set of each n-channel combination is predicted, and the second prediction accuracy of each n-channel combination is obtained. The n channels in the n-channel combination with the highest prediction accuracy are selected as the n preset channels.

Images