TWI806079B - Medication inventory method, program, and readable medium of emergency vehicle - Google Patents

Abstract

An emergency vehicle having medication inventory system as well as a system, a method, a program, and a readable medium of medication inventory are disclosed. The method includes: acquiring an image data of an image of a medicine stored in an emergency vehicle; using an AI image identification model to identify a name and a quantity of the medicine; using a processing device to compare the identified name and quantity of the medicine with a preset medicine name and a preset medicine quantity, respectively; the processing device controlling an output device to output a first reminder while the identified name of the medicine is inconsistent with the preset medicine name (or, while the identified quantity of the medicine is inconsistent with the preset medicine quantity). According to the present disclosure, names and quantities of medicines can be identified only according to their respective image data. There is no need to scan the medicines one by one. Moreover, a first reminder is raised while the identified name or quantity of medicine is abnormal so that a user can complete the medication inventory quickly, thereby grasping the medication storage status of the emergency vehicle.

Description

Medicine inventory method, program and readable media in ambulance vehicles

The invention relates to a method, a program and a readable medium for counting medicines in an ambulance.

Once a patient has an unexpected first aid, the integrity and function of the emergency trolley equipment may affect the timeliness of the patient's first aid. The Hospital Evaluation and Medical Quality Promotion Association is also a hospital evaluation benchmark and evaluation item Article 2.3.13 proposes: first-aid equipment and medicines should have clear management regulations to ensure equipment functions and medicines and health materials, and to ensure correct operation. In addition, most ambulance medicines are high-alert medicines. How to improve drug safety is also one of the International Patient Safety Goals (IPSG).

In order to effectively manage the ambulances, traditionally, the nursing staff divided the ambulances into three shifts, and recorded the shift results of each drug on the paper audit form one by one. Not only was the shift time-consuming, but the paper audit form It is also more difficult to preserve.

Therefore, the Republic of China Patent Publication No. M577582 provides a mobile medical material management system, which may include electronic devices and management servers. The electronic device includes a scanner and a display interface. The management server stores the management information of a plurality of medical materials, and the management information of each medical material includes product name, batch number, expiry date and inventory quantity. Among them, each medical material contains a medical material barcode, and the medical material barcode corresponds to the product name, batch number and expiration date of the medical material. When the scanner scans the medical material barcode of the medical material, the display interface displays The product name, batch number and expiry date of the medical material, and the management server finds the stock quantity of the medical material according to the barcode of the medical material, and displays the stock quantity of the medical material through the display interface.

However, in the aforementioned patent case, the nursing staff has to turn over the barcodes of the medical materials one by one to an angle that can be scanned by the scanner, and then scan each barcode of the medical materials one by one, which is still very inconvenient to operate.

Therefore, the present inventor proposes an aided inventory system for emergency vehicle medicine image recognition, which is used to count a medicine in an emergency vehicle. Image recognition model; an image capture unit, signal connected to the processing unit, the image capture unit corresponds to a drug position of the ambulance, and the drug is stored in the drug position; an output unit, signal connected to the processing unit; through The image capture unit obtains an image data of the location of the drug in the ambulance and sends it to the artificial intelligence image recognition model, and the image data includes the image of the drug, and the artificial intelligence image recognition model The image of a drug item on the drug position and a corresponding drug quantity are identified, and the processing unit then compares a drug default item corresponding to the drug position and a drug quantity according to the drug item and the drug quantity. The preset quantity of the medicine, when the item of the medicine does not match the preset item of the medicine, or the quantity of the medicine does not match the preset quantity of the medicine, the processing unit controls the output unit to output a first prompt.

Further, there is a weight sensing unit signally connected to the processing unit, the medicine position is a drawer, the weight sensing unit is set in the ambulance and corresponds to the drawer, or, the drawer has a plurality of storage compartments, and there are multiple weights The sensing unit is arranged in the ambulance and corresponds to each storage compartment respectively; the weight sensing unit obtains a drug weight of the drawer or each storage compartment and sends it to the processing unit, and the processing unit then according to the Item and a drug preset weight of the drug quantity, compared with the drug When the weight and the preset weight of the medicine are not matched, the processing unit controls the output unit to output a second prompt.

Further, the processing unit performs in-depth learning to establish the artificial intelligence image recognition model based on the plurality of training image data and an actual drug item and an actual drug quantity corresponding to each training image data.

The inventor also proposes an aided counting method of drug image recognition in an ambulance, which includes: obtaining an image data of a drug location in an ambulance, and the image data includes an image of a drug; inputting the image data into an artificial intelligence image recognition Model, the artificial intelligence image recognition model recognizes a drug item and a corresponding drug quantity of the drug at the position of the drug according to the image of the drug in the image data; input the drug item and the drug quantity into a process unit, the processing unit compares a drug preset item corresponding to the drug position with the drug item, and a drug preset quantity with the drug quantity; when the drug item does not match the drug preset item, or When the quantity of the medicine does not match the preset quantity of the medicine, the processing unit controls an output unit to output a first prompt.

Further, a medicine weight of the medicine position of the ambulance is obtained and input into the processing unit, and the processing unit obtains a medicine preset weight corresponding to the medicine item and the medicine quantity, and compares the medicine weight and the medicine preset weight. The weight is set, and when the comparison is inconsistent, the processing unit controls the output unit to output a second prompt.

Further, when the drug item matches the drug preset item, the drug quantity matches the drug preset quantity, or, after the processing unit controls the output unit to output the first prompt, the processing unit item, the preset item of the drug, and the comparison result of the quantity of the medicine and the preset quantity of the medicine, sorted into a shift record and sent to the output unit for output, and stored in a database.

Further, after the output unit outputs the shift record, it obtains an updated image data and a report information of the ambulance, and the artificial intelligence image recognition model recognizes an updated drug item and an updated drug quantity input according to the updated image data The processing unit, the processing unit then arranges the updated drug item, the drug default item and the updated drug quantity, the drug preset quantity comparison result, and the report information as a supplementary shift record and sends it to the The output unit outputs and stores in the database.

The present inventor also proposes a computer program and a computer readable medium, the computer readable medium stores the computer program, and is used to implement the above-mentioned method for assisting the inventory of medicines in emergency vehicles after being installed in a computer.

The inventor further proposes an ambulance with a drug image recognition auxiliary inventory system for counting a drug. The emergency vehicle with a drug image recognition auxiliary inventory system includes: an ambulance with a drug location, and the drug is stored in the The location of the drug; a processing unit with a pre-trained artificial intelligence image recognition model; an image capture unit connected to the processing unit with a signal corresponding to the location of the drug; and an output unit connected to the processing unit with a signal Obtain an image data of the position of the drug in the ambulance through the image capture unit and send it to the artificial intelligence image recognition model, and the image data includes the image of the drug, and the artificial intelligence image recognition model The image of the drug identifies a drug item at the drug position and a corresponding drug quantity, and the processing unit compares a drug preset item corresponding to the drug position according to the drug item and the drug quantity and a preset quantity of medicine, when the item of the medicine does not match the preset item of the medicine, or the quantity of the medicine does not match the preset quantity of the medicine, the processing unit controls the output unit to output a first prompt.

According to the above-mentioned technical characteristics, the following effects can be preferably achieved:

1. With the artificial intelligence image recognition model, the quantity of each drug item can be identified from a single image data without scanning the drugs one by one, and the first reminder will be issued when the drug item or drug quantity is abnormal, which is convenient for users Complete the order quickly and master the status of the medicine in the ambulance.

2. In addition to the judgment of image data, coupled with the assistance of the weight of medicines, the ordering of medicines can have a higher accuracy rate.

3. In addition to being sent to the output unit for output in real time, the shift record and make-up record can also be stored in the database, which is convenient for data preservation and future audit and query.

1: Processing unit

11: AI image recognition model

2: Image capture unit

3: Output unit

4: Weight sensing unit

5: Database

A: Ambulance

A1: drawer

A2: cell

B: medicine

C: user

[The first figure] is a schematic flow chart of the embodiment of the present invention.

[The second figure] is a system block diagram of the embodiment of the present invention.

[Figure 3] is the implementation schematic diagram 1 of the embodiment of the present invention, which shows that the image data is obtained by the image capture unit.

[Figure 4] is the second implementation schematic diagram of the embodiment of the present invention, showing image data.

[FIG. 5] is the third implementation schematic diagram of the embodiment of the present invention, showing the preparation for the implementation of the aided counting method of the emergency vehicle drug image recognition.

[The sixth figure] is the implementation schematic diagram 4 of the embodiment of the present invention, showing shift records.

[Fig. 7] is the implementation schematic diagram 5 of the embodiment of the present invention, indicating the input of return information.

[Picture 8] is the sixth implementation schematic diagram of the embodiment of the present invention, showing the confirmation of the return information.

[The ninth figure] is the implementation schematic diagram seven of the embodiment of the present invention, indicating the record of making up shifts.

[Figure 10] is the eighth implementation schematic diagram of the embodiment of the present invention, showing the shift statistics table obtained according to the shift records.

[Eleventh figure] is the implementation schematic diagram 9 of the embodiment of the present invention, showing the shift statistics table obtained according to the shift records.

Based on the above technical features, the main functions of the ambulance drug counting method, program and readable media of the present invention will be clearly presented in the following embodiments.

Please refer to the first figure to the third figure, which disclose the medical image recognition auxiliary inventory system of the ambulance according to the embodiment of the present invention, which is installed in an ambulance A to become one of the emergency vehicles with the medical image recognition auxiliary inventory system of the present invention. The emergency vehicle drug image recognition auxiliary inventory system is used to count a drug B in the ambulance A [please match the drug B with the fourth picture] according to an emergency vehicle drug image recognition auxiliary inventory method of the present invention. In actual implementation, a computer program, or a computer-readable medium storing the computer program, can also be installed in a computer to execute the method for assisting the counting of medicines in emergency vehicles through image recognition.

The medical image recognition auxiliary inventory system for ambulances includes: a processing unit 1 with a pre-trained artificial intelligence image recognition model 11 .

More specifically, the processing unit 1 performs deep learning to establish the artificial intelligence image recognition model 11 based on the plurality of training image data and an actual drug item and an actual drug quantity corresponding to each training image data. For example, the algorithm for the processing unit 1 to establish the artificial intelligence image recognition model 11 can adopt one of the following algorithms: Random Forest algorithm (Random Forest), Support Vector Machines (Support Vector Machines, SVM), K-neighborhood Algorithm (K Nearest Neighbor, KNN), multilayer perceptron (Multilayer Perceptron, MLP), lightweight gradient boosting model (Light Gradient Boosting Machine, LightGBM), extreme gradient boosting (eXtreme Gradient Boosting, XGBoost) and logistic regression analysis ( Logistic Regression).

An image capture unit 2 is signally connected to the processing unit 1 . The image capture unit 2 corresponds to a medicine position of the ambulance A, and the medicine B is stored in the medicine position, preferably, the medicine position is a drawer A1 of the ambulance A. The image capture unit 2 can be fixed on the ambulance A, or as shown in the third figure, a user C holds the image capture unit 2 towards the drawer A1.

An output unit 3 is connected to the processing unit 1 for signals. Preferably, the output unit 3 can be the mobile phone or tablet of the user C, and the image capture unit 2 is integrated on the output unit 3 .

A weight sensing unit 4 is signal-connected to the processing unit 1 , the weight sensing unit 4 is set in the ambulance A and corresponds to the drawer A1 .

A database 5 is signally connected to the processing unit 1 .

Please refer to the third figure to the fifth figure, and please match the second figure. Preferably, the drawer A1 is separated from a plurality of storage compartments A2, and a plurality of the weight sensing units 4 are arranged on the ambulance A and correspond to Each cell A2. Preferably, both the weight sensing unit 4 and the storage compartment A2 are inside the drawer A1, and the weight sensing unit 4 is combined under the storage compartment A2.

When starting to execute the aided counting method of the ambulance drug image recognition, since the output unit 3 is the mobile phone or tablet of the user C, the user C can directly select the shift time on the output unit 3 .

Then, the user C obtains an image data of the drug position of the ambulance A through the image capture unit 2, as shown in the fourth figure, and the image data includes the image of the drug B, and the image capture The unit 2 sends the image data to the artificial intelligence image recognition model 11 .

The artificial intelligence image recognition model 11 recognizes a drug item at the position of the drug and a corresponding drug quantity according to the image of the drug B in the image data.

The processing unit 1 then compares a preset drug item corresponding to the drug position and a preset drug quantity according to the drug item and the drug quantity, and when the drug item does not match the drug preset item, or When the quantity of the medicine does not match the preset quantity of the medicine, the processing unit 1 controls the output unit 3 to output a first prompt.

At the same time, the weight sensing unit 4 obtains a drug weight of the drawer A1 or each storage compartment A2 and sends it to the processing unit 1, and the processing unit 1 then calculates the weight of a drug according to the drug item and the drug quantity. Set the weight, compare the weight of the drug with the preset weight of the drug, and when the comparison is inconsistent, the processing unit 1 controls the output unit 3 to output a second prompt.

In actual implementation, the first prompt and the second prompt may also be issued in combination.

It should be particularly noted that the processing unit 1 first recognizes the drug item on the drug position and the corresponding drug quantity based on the image of the drug B in the image data, and compares the drug item corresponding to the drug position. After the preset item of the medicine and the preset quantity of the medicine are compared, the weight of the medicine and the preset weight of the medicine are compared.

Once the item of the drug does not match the preset item of the drug, or the quantity of the drug does not match the preset quantity of the drug, the processing unit 1 will not compare the weight of the drug with the preset weight of the drug, only the item matches the preset item of the medicine, and the quantity of the medicine matches the preset quantity of the medicine, the processing unit 1 will compare the weight of the medicine with the preset weight of the medicine.

Please refer to the first picture, the second picture and the sixth picture, when the drug item and the drug preset item, the drug quantity and the drug preset quantity, the drug weight and the drug preset weight all match, or After the processing unit 1 controls the output unit 3 to output the first prompt and/or the second prompt, the processing unit 1 sets the drug item, the drug preset item, the drug quantity, and the drug preset The comparison results of the quantity are sorted into one-point shift records and sent to the output unit 3 for output, and are stored in the database 5 .

Take Solu-cortef (cortisol) in the fifth row as an example, the shift record shows 1/3, 1 is the quantity of the medicine, and 3 is the preset quantity of the medicine, that is to say, Solu-cortef 2 are missing. However, the remaining quantity of the drug and the preset quantity of the drug in the sixth figure are only examples and may not be related to other figures.

Please refer to the second picture, the third picture and the seventh picture, and please match the first picture. After the output unit 3 outputs the shift record, because there are two Solu-cortefs missing, the user C performs Solu-cortef supplementary classes.

The user C reacquires an updated image data of the ambulance A after the make-up shift is completed, and inputs a report information to the processing unit 1, for example, the Solu-cortef is missing due to offset, patient CPR or other reasons .

Since the output unit 3 is the mobile phone or tablet of the user C, the output unit 3 can also have the function of inputting the report information, and when the output unit 3 is only a display, an additional input unit ( not shown) to enter the report information.

Please refer to the second figure, the third figure and the eighth figure, after the user C inputs the report information, the processing unit 1 can re-display multiple pieces of the report information to the user C for confirmation and send them out.

Please refer to the second picture, the third picture and the ninth picture, and please match the first picture, the updated image data is sent to the artificial intelligence image recognition model 11, and the artificial intelligence image recognition model 11 recognizes the image according to the updated image data Output an updated drug item and an updated drug quantity, and input them into the processing unit 1, and the processing unit 1 then compares the updated drug item, the preset drug item with the updated drug quantity, and the drug preset quantity The result and the report information are sorted into a supplementary shift record, sent to the output unit 3 for output, and stored in the database 5 .

Only after the processing unit 1 obtains the updated image data and the report information, the processing unit 1 will output the make-up shift record.

Please refer to the second picture, the tenth picture and the eleventh picture, the processing unit 1 sorts out the shift record and/or the make-up shift record of multiple time periods, and can also be further sorted into a one-time shift statistics table and sent to the The output unit 3 outputs and stores in the database 5 . For example, the processing unit 1 can compare this week with last week, this month with last month, even this year with last year, etc., as the evaluation standard and performance check of the hospital.

Please refer to the second to fourth figures again. With the artificial intelligence image recognition model 11, without scanning the medicine B one by one, the quantity of the medicine of each medicine item can be identified from the single image data, and the When the drug item or the quantity of the drug is abnormal, a first prompt is issued, so that the user C can quickly complete the shift and grasp the status of the drug B in the ambulance A.

In addition to the judgment of the image data, coupled with the assistance of the weight of the drug, the ordering of the drug B can have a higher accuracy rate. In addition, besides being sent to the output unit 3 for output in real time, the shift count record and the supplementary shift record can also be stored in the database 5, which is convenient for data preservation and future audit and query.

Based on the description of the above-mentioned embodiments, it is possible to fully understand the operation of the present invention, use and the effect that the present invention produces, but the above-mentioned embodiments are only preferred embodiments of the present invention, and should not be used to limit the implementation of the present invention. The scope, that is, the simple equivalent changes and modifications made according to the patent scope of the present invention and the content of the description of the invention, all fall within the scope of the present invention.

Claims (5)

A method for assisting drug image recognition in an emergency vehicle, comprising: obtaining an image data of a drug location in an ambulance, and the image data includes an image of a drug; inputting the image data into an artificial intelligence image recognition model, the artificial intelligence The image recognition model identifies a drug item and a corresponding drug quantity of the drug at the position of the drug according to the image of the drug in the image data; inputs the drug item and the drug quantity into a processing unit, and the processing unit Compare a preset drug item corresponding to the drug position with the drug item, and a drug preset quantity with the drug quantity; obtain a drug weight at the drug position of the ambulance and input it into the processing unit, the processing The unit obtains a preset medicine weight corresponding to the medicine item and the medicine quantity; when the medicine item does not match the medicine preset item, or the medicine quantity does not match the medicine preset quantity, the processing unit controls An output unit outputs a first reminder; when the item of the medicine matches the preset item of the medicine, and the quantity of the medicine matches the preset quantity of the medicine, the processing unit compares the weight of the medicine with the preset quantity of the medicine. Weight, when the comparison does not match, the processing unit controls the output unit to output a second prompt; obtains an updated image data and a report information of the ambulance, and the artificial intelligence image recognition model recognizes an updated image based on the updated image data The drug item and an updated drug quantity are input into the processing unit, and the processing unit then compares the updated drug item, the drug default item with the updated drug quantity, the drug default quantity, and the return information Organize it into a supplementary shift record and send it to the output unit for output. As described in Claim 1, the image recognition-assisted inventory method for emergency vehicles, further, when the item of the drug matches the preset item of the drug, and the quantity of the drug matches the preset quantity of the drug, or the processing unit controls the After the output unit outputs the first prompt, the processing unit The preset item of the product, and the comparison result of the quantity of the medicine and the preset quantity of the medicine are sorted into a shift record and sent to the output unit for output, and stored in a database. As described in claim item 2, the image recognition auxiliary inventory method for ambulance medicines, further, the output unit outputs the shift record before obtaining the updated image data and the report information of the ambulance, and storing the supplementary shift record in the database. A computer program, which is used to execute the image recognition-assisted counting method for medicines in emergency vehicles as described in any one of Claims 1 to 3 after being installed in a computer. A computer-readable medium storing a computer program for executing the image recognition-assisted counting method for medicines in emergency vehicles as described in any one of claims 1 to 3 after being installed in a computer.

Images