CN111382622A - Medicine identification system based on deep learning and implementation method thereof - Google Patents
Medicine identification system based on deep learning and implementation method thereof Download PDFInfo
- Publication number
- CN111382622A CN111382622A CN201811628692.9A CN201811628692A CN111382622A CN 111382622 A CN111382622 A CN 111382622A CN 201811628692 A CN201811628692 A CN 201811628692A CN 111382622 A CN111382622 A CN 111382622A
- Authority
- CN
- China
- Prior art keywords
- medicine
- module
- medicines
- information
- algorithm
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H20/00—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
- G16H20/10—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to drugs or medications, e.g. for ensuring correct administration to patients
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medicinal Chemistry (AREA)
- Epidemiology (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Physics & Mathematics (AREA)
- Multimedia (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Primary Health Care (AREA)
- Public Health (AREA)
- Image Analysis (AREA)
- Medical Preparation Storing Or Oral Administration Devices (AREA)
Abstract
The invention discloses a medicine identification system based on deep learning and an implementation method thereof, wherein the medicine identification system comprises a picture acquisition module for acquiring medicine picture information; the detection and segmentation module is used for calculating the accurate position of each medicine and segmenting the medicines one by one; the controller is used for receiving the feedback information, receiving the comparison information, receiving the medicine list information and sending a control instruction; the identification module is used for identifying the information of characters, texts and pictures of the medicine box and accurately identifying the types of the medicines and the number of each type; and the comparison module is used for comparing the medicine information with the medicine list of the patient. The medicine identification system based on deep learning and the implementation method thereof can automatically and accurately identify the types and the quantity of medicines, improve the medicine placing speed of a pharmacy, replace the manual operation process, reduce the labor cost and reduce the probability of errors caused by manual medicine distribution.
Description
Technical Field
The invention relates to the technical field of medicine identification, in particular to a medicine identification system based on deep learning and an implementation method thereof.
Background
With the rapid development of the pharmaceutical industry, the medicines on the market are various, and most of the medicines in hospitals are thousands of medicines, wherein different specifications and different dosages of the same medicine have various forms.
In hospitals still using manual medicine picking and dispensing, hospital medicine taking windows and before finally delivering medicines to patients, a worker is needed to confirm whether the types and the quantities of the medicines are correct or not, the medicine dispensing speed is low, the situations that the worker misreads a prescription transmitted by a Hospital Information System (HIS) and the medicines are similar in package or easily confused and add errors exist, and the like exist, and the workload of pharmacists and the probability of medicine dispensing errors are increased.
Therefore, a medicine identification system based on deep learning and an implementation method thereof are provided.
Disclosure of Invention
The invention aims to provide a medicine identification system based on deep learning and an implementation method thereof, which can automatically and accurately identify the type and the quantity of medicines, improve the medicine dispensing speed of a pharmacy and reduce the labor cost.
In order to achieve the purpose, the invention provides the following technical scheme:
a medicine identification system based on deep learning and a realization method thereof comprise a medicine identification system, and are characterized in that: the drug identification system comprises:
the picture acquisition module is used for acquiring medicine picture information;
the detection and segmentation module is used for calculating the accurate position of each medicine and segmenting the medicines one by one;
the controller is used for receiving the feedback information, receiving the comparison information, receiving the medicine list information and sending a control instruction;
the identification module is used for identifying the information of characters, texts and pictures of the medicine box and accurately identifying the types of the medicines and the number of each type;
the comparison module is used for comparing the medicine information with the medicine list of the patient;
the feedback module is used for feeding back the pictures, names and the number of the medicines which are not successfully paired;
a hospital drug inventory system for providing an inventory of patients' drugs;
the communication module is used for communicating the controller with the hospital drug inventory system;
the display screen is used for displaying the information of the medicine comparison result;
the output end of the picture acquisition module is connected with the input end of the detection and segmentation module, the output end of the detection and segmentation module is connected with the input end of the controller, and the controller is in bidirectional signal connection with a hospital drug list system through a communication module;
the output of controller is connected with identification module's input, and the controller with compare module both-way signal connection, and compare the input of module and be connected with identification module's output, should compare the output of module and be connected with the feedback input of controller through feedback module, two outputs of controller are connected with two inputs of display screen respectively.
An implementation method of a medicine identification system based on deep learning comprises the following steps:
s1: the staff puts the medicine (A, B, C, D, etc.) of a certain (A) into the medicine frame;
s2: the staff puts the medicine frame into the equipment through the medicine taking port of the equipment;
s3: the picture acquisition module shoots pictures of the medicines in the medicine frame and sends the pictures to the artificial intelligence algorithm module;
s4: the artificial intelligence algorithm module identifies the types of the medicines and outputs medicine information comprising the types of the medicines and the quantity of each medicine;
s5: the controller is in butt joint with a medicine list system of a hospital through the communication module to obtain a medicine list of a certain medicine (A);
s6: the drug information output by the S4 is matched and checked with the drug list of a certain drug (A) one by one;
s7: the display displays the type and the quantity of the certain medicine (A) in the medicine frame and the matching result of the certain medicine (A) in the medicine frame and the medicine list, if the medicines are completely matched, the medicine comparison success and the medicine information are output, and if the medicines are not completely matched, the pictures, the names and the quantity of the medicines which are not successfully matched are output.
Preferably, the controller is a CPU.
Preferably, the image acquisition module adopts a camera.
Preferably, the recognition module includes a text detection module and a character recognition module, the text detection module adopts EAST natural scene text detection algorithm, and the character recognition module adopts CRNN algorithm.
Preferably, the detection segmentation module adopts a MaskRcnn example segmentation algorithm.
Preferably, the artificial intelligence algorithm in S4 includes a model training algorithm and a recognition algorithm;
the model training algorithm comprises:
a1: obtaining medicine pictures in the medicine blue in an actual scene through a camera, and arranging to obtain a picture library;
a2: marking the medicines in the medicine picture library one by one, wherein the medicines comprise the position, the type, the medicines and the like of each medicine, and dividing a data set into a training set, a verification set and a test set;
a3: establishing a deep learning algorithm model required by a medicine identification task through a research problem;
a4: training an algorithm model by using a training set and a verification set, and iteratively updating parameters at each time until the model converges;
a5: further testing the performance of the model by using the test set, and stopping training if the performance meets the requirements; if the model performance does not meet the requirements, repeating the steps by increasing the data volume and modifying the deep learning algorithm model until the model performance meets the requirements;
the recognition algorithm comprises:
b1 picture acquisition module: shooting a medicine picture by using a camera, and sending the medicine picture to a detection segmentation module;
b2 detection segmentation module: calculating the accurate position of each medicine by using an artificial intelligence algorithm, and dividing the medicines one by one;
b3 identifies the module: the information such as the shape, characters, trademarks, pictures and the like of the medicine box is identified by an artificial intelligence algorithm, and the types of the medicines and the number of each type are accurately identified;
b4 alignment module: comparing and checking the output medicine information with the medicine list of the patient, and if the quantity and the type can be matched, passing; if the comparison is not successful, the pictures, names and number of the medicines which are not successfully matched are measured and output.
Compared with the prior art, the invention has the beneficial effects that: the medicine identification system based on deep learning and the implementation method thereof can automatically and accurately identify the types and the quantity of medicines, improve the medicine placing speed of a pharmacy, replace the manual operation process, reduce the labor cost and reduce the probability of errors caused by manual medicine distribution.
Drawings
FIG. 1 is a schematic diagram of a system for drug identification according to the present invention;
FIG. 2 is a block diagram of the identification unit of the present invention;
FIG. 3 is a model diagram of a MaskRcnn example segmentation algorithm of the present invention;
FIG. 4 is a model diagram of an EAST natural scene text detection algorithm of the present invention;
FIG. 5 is a diagram of a model of the CRNN algorithm of the present invention;
FIG. 6 is a flow chart of an artificial intelligence algorithm of the present invention;
FIG. 7 is an exploded front view of the drug identification device of the present invention;
fig. 8 is an exploded view of the back structure of the drug identification device of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1-8, the present invention provides a technical solution: a medicine identification system based on deep learning and an implementation method thereof comprise a medicine identification system, wherein the medicine identification system comprises:
the picture acquisition module is used for acquiring medicine picture information;
the detection and segmentation module is used for calculating the accurate position of each medicine and segmenting the medicines one by one;
the controller is used for receiving the feedback information, receiving the comparison information, receiving the medicine list information and sending a control instruction;
the identification module is used for identifying the information of characters, texts and pictures of the medicine box and accurately identifying the types of the medicines and the number of each type;
the comparison module is used for comparing the medicine information with the medicine list of the patient;
the feedback module is used for feeding back the pictures, names and the number of the medicines which are not successfully paired;
a hospital drug inventory system for providing an inventory of patients' drugs;
the communication module is used for communicating the controller with the hospital drug inventory system;
the display screen is used for displaying the information of the medicine comparison result;
the output end of the picture acquisition module is connected with the input end of the detection and segmentation module, the output end of the detection and segmentation module is connected with the input end of the controller, and the controller is in bidirectional signal connection with a hospital drug list system through a communication module;
the output of controller is connected with identification module's input, and the controller with compare module both-way signal connection, and compare the input of module and be connected with identification module's output, should compare the output of module and be connected with the feedback input of controller through feedback module, two outputs of controller are connected with two inputs of display screen respectively.
An implementation method of a medicine identification system based on deep learning comprises the following steps:
s1: the staff puts the medicine (A, B, C, D, etc.) of a certain (A) into the medicine frame;
s2: the staff puts the medicine frame into the equipment through the medicine taking port of the equipment;
s3: the picture acquisition module shoots pictures of the medicines in the medicine frame and sends the pictures to the artificial intelligence algorithm module;
s4: the artificial intelligence algorithm module identifies the types of the medicines and outputs medicine information comprising the types of the medicines and the quantity of each medicine;
s5: the controller is in butt joint with a medicine list system of a hospital through the communication module to obtain a medicine list of a certain medicine (A);
s6: the drug information output by the S4 is matched and checked with the drug list of a certain drug (A) one by one;
s7: the display displays the type and the quantity of the certain medicine (A) in the medicine frame and the matching result of the certain medicine (A) in the medicine frame and the medicine list, if the medicines are completely matched, the medicine comparison success and the medicine information are output, and if the medicines are not completely matched, the pictures, the names and the quantity of the medicines which are not successfully matched are output.
In the invention: the controller employs a CPU.
In the invention: the image acquisition module adopts a camera.
In the invention: the recognition module comprises a text detection module and a character recognition module, the text detection module adopts an EAST natural scene text detection algorithm, and the character recognition module adopts a CRNN algorithm.
In the invention: the detection segmentation module adopts a MaskRcnn example segmentation algorithm.
In the invention: the artificial intelligence algorithm in the S4 comprises a model training algorithm and a recognition algorithm;
the model training algorithm comprises:
a1: obtaining medicine pictures in the medicine blue in an actual scene through a camera, and arranging to obtain a picture library;
a2: marking the medicines in the medicine picture library one by one, wherein the medicines comprise the position, the type, the medicines and the like of each medicine, and dividing a data set into a training set, a verification set and a test set;
a3: establishing a deep learning algorithm model required by a medicine identification task through a research problem;
a4: training an algorithm model by using a training set and a verification set, and iteratively updating parameters at each time until the model converges;
a5: further testing the performance of the model by using the test set, and stopping training if the performance meets the requirements; if the model performance does not meet the requirements, repeating the steps by increasing the data volume and modifying the deep learning algorithm model until the model performance meets the requirements;
the recognition algorithm comprises:
b1 picture acquisition module: shooting a medicine picture by using a camera, and sending the medicine picture to a detection segmentation module;
b2 detection segmentation module: calculating the accurate position of each medicine by using an artificial intelligence algorithm, and dividing the medicines one by one;
b3 identifies the module: the information such as the shape, characters, trademarks, pictures and the like of the medicine box is identified by an artificial intelligence algorithm, and the types of the medicines and the number of each type are accurately identified;
b4 alignment module: comparing and checking the output medicine information with the medicine list of the patient, and if the quantity and the type can be matched, passing; if the comparison is not successful, the pictures, names and number of the medicines which are not successfully matched are measured and output.
In the invention: the detection segmentation module adopts a MaskRcnn example segmentation algorithm; the method comprises the following specific steps: please refer to fig. 3: MaskRcnn example segmentation algorithm:
model training
1) Marking the medicine pictures shot by the camera, wherein the medicine pictures comprise the positions of the medicines and the types of the medicines and are used as labels for training of the segmentation model.
2) A tensierflow deep learning framework is used to build a network structure model as shown in the following figure. The inputs to the network include: pictures, and labels for pictures. Wherein the label comprises a picture of a split label for the so-called location of each drug, a bounding box for each drug, the type of drug (here only one, if not a drug). The output of the network model includes: rpn network output (including the position and type of the Proposal), branch split output (including the position of the pixel occupied by each drug split by the network, which is a binary image), branch detection output (including the position of the border of each drug detected by the network, which is a rectangular box formed by four coordinates), and network classification output (including the type of drug, which is the probability of being a drug or not).
3) Sending the data in the step (1) to the network in the step (2) to obtain data, and calculating a loss function, wherein the loss function is as follows:
L＝Lcls+Lbox+LmaskL＝Lcls+Lbox+Lmask
lcs-class loss
Lbox-bounding-box regression loss
Lmak-mask partition loss
4) And continuously and iteratively updating the network by using the BP algorithm so as to converge the network.
5) Evaluating performance of a network using a test set
Prediction of model
1) After the medicine is placed in the device, the camera is used for taking a picture to obtain a picture of the medicine.
2) And sending the medicine pictures into the trained segmentation network to obtain the segmentation result of the medicine.
3) And (4) using a minimum external quadrilateral frame for the segmentation result (a binary image), and rightly cutting the quadrilateral to obtain a picture of each medicine.
In the invention: the recognition module comprises a text detection module and a character recognition module; the text detection module adopts EAST natural scene text detection algorithm.
Please refer to fig. 4: EAST natural scene text detection algorithm
Model training part
1) A data set is prepared, which contains two parts, one part being an open data set (including an ICDAR data set, an ICPR data set of a sky pool) and the other part being a self-labeling data set.
2) A tensierflow deep learning framework is used to build a network structure model as shown in the following figure. The input to the network includes pictures and labels, where the labels include a score map, which represents the probability of whether each point is literal, and a geometry map, which represents the location of the point.
3) Inputting the data in (1) into the network in (2) to obtain data, and calculating a loss function, wherein the loss function is as follows:
L＝Ls+λgLg
where Ls and Lg represent the loss of score map and geometry map, respectively, λgThe weight representing the two losses is set to 1 here.
4) And continuously and iteratively updating the network by using the BP algorithm so as to converge the network.
5) Evaluating performance of a network using a test set
Prediction part of model
1) And sending the pictures of the medicines obtained by the division module into the model to obtain a socre map and a geometry map.
2) Setting a threshold value, and converting the score map and the geometry map into a text box and a score of the text box.
3) And filtering redundant text boxes by using a Locality-Aware NMS algorithm to obtain a final detection result.
In the invention: the character recognition module adopts a CRNN algorithm, which specifically comprises the following steps:
please refer to fig. 5: the CRNN algorithm:
an algorithm training part:
1) a data set is prepared, which includes two parts, a common text data set and a drug name data set, and the drug name data set is created by first collecting a list of drug names and then expanding the data set by using different fonts and backgrounds.
2) A network structure model as shown in the following figure was built using tensoflow. And inputting the data into the network, and continuously iterating through a BP algorithm to make the network converge.
3) Evaluating performance of a network using a test set
Network prediction part
1) Cutting out the text box detected by EAST and correcting
2) Sorting the height of the detection frames from big to small
3) And (5) sending the high resize of the detection frame to 32, sending the high resize to the network to obtain a predicted result vector, and translating the characters by referring to the dictionary.
4) And comparing the recognized characters with the medical note information of the patient, and if the name of the medicine can be matched, outputting the name of the medicine and the position of the text box in the original image. If none of the matching is found, the medicine is output as the medicine which is not found on the bill.
In the invention: the structure of the medication identification device is illustrated in FIGS. 7-8;
the medicine identification device comprises a shell, the shell is of a hollow structure, a medicine taking port is formed in the position, close to the bottom, of the front surface of the shell, a display screen is fixedly mounted on the position, close to the middle, of the front surface of the shell, a power supply and a controller are fixedly connected in sequence from left to right in the middle of the shell, the bottom of the controller is fixedly connected with an interface board, the bottom of the interface board is fixedly connected with a diffusion plate, an interface is fixedly mounted at the top of the back of the shell, an upper cover is fixedly connected with the top of the shell, a lens is fixedly mounted at the position, close to the middle, of the bottom of the upper cover, and the bottom of the lens penetrates.
In summary, the following steps: the medicine identification system based on deep learning and the implementation method thereof can automatically and accurately identify the types and the quantity of medicines, improve the medicine placing speed of a pharmacy, replace the manual operation process, reduce the labor cost and reduce the probability of errors caused by manual medicine distribution.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
The related modules involved in the system are all hardware system modules or functional modules combining computer software programs or protocols with hardware in the prior art, and the computer software programs or the protocols involved in the functional modules are all known in the technology of persons skilled in the art, and are not improvements of the system; the improvement of the system is the interaction relation or the connection relation among all the modules, namely the integral structure of the system is improved, so as to solve the corresponding technical problems to be solved by the system.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (7)
1. A medicine identification system based on deep learning and a realization method thereof comprise a medicine identification system, and are characterized in that: the drug identification system comprises:
the picture acquisition module is used for acquiring medicine picture information;
the detection and segmentation module is used for calculating the accurate position of each medicine and segmenting the medicines one by one;
the controller is used for receiving the feedback information, receiving the comparison information, receiving the medicine list information and sending a control instruction;
the identification module is used for identifying the information of characters, texts and pictures of the medicine box and accurately identifying the types of the medicines and the number of each type;
the comparison module is used for comparing the medicine information with the medicine list of the patient;
the feedback module is used for feeding back the pictures, names and the number of the medicines which are not successfully paired;
a hospital drug inventory system for providing an inventory of patients' drugs;
the communication module is used for communicating the controller with the hospital drug inventory system;
the display screen is used for displaying the information of the medicine comparison result;
the output end of the picture acquisition module is connected with the input end of the detection and segmentation module, the output end of the detection and segmentation module is connected with the input end of the controller, and the controller is in bidirectional signal connection with a hospital drug list system through a communication module;
the output of controller is connected with identification module's input, and the controller with compare module both-way signal connection, and compare the input of module and be connected with identification module's output, should compare the output of module and be connected with the feedback input of controller through feedback module, two outputs of controller are connected with two inputs of display screen respectively.
2. A method for realizing a medicine identification system based on deep learning is characterized in that: the implementation method of the medicine identification system comprises the following steps:
s1: the staff puts the medicine (A, B, C, D, etc.) of a certain (A) into the medicine frame;
s2: the staff puts the medicine frame into the equipment through the medicine taking port of the equipment;
s3: the picture acquisition module shoots pictures of the medicines in the medicine frame and sends the pictures to the artificial intelligence algorithm module;
s4: the artificial intelligence algorithm module identifies the types of the medicines and outputs medicine information comprising the types of the medicines and the quantity of each medicine;
s5: the controller is in butt joint with a medicine list system of a hospital through the communication module to obtain a medicine list of a certain medicine (A);
s6: the drug information output by the S4 is matched and checked with the drug list of a certain drug (A) one by one;
s7: the display displays the type and the quantity of the certain medicine (A) in the medicine frame and the matching result of the certain medicine (A) in the medicine frame and the medicine list, if the medicines are completely matched, the medicine comparison success and the medicine information are output, and if the medicines are not completely matched, the pictures, the names and the quantity of the medicines which are not successfully matched are output.
3. The drug identification system and the realization method thereof based on deep learning of claim 1, wherein: the controller adopts a CPU.
4. The drug identification system and the realization method thereof based on deep learning of claim 1, wherein: the image acquisition module adopts a camera.
5. The drug identification system and the realization method thereof based on deep learning of claim 1, wherein: the recognition module comprises a text detection module and a character recognition module, the text detection module adopts an EAST natural scene text detection algorithm, and the character recognition module adopts a CRNN algorithm.
6. The drug identification system and the realization method thereof based on deep learning of claim 1, wherein: the detection segmentation module adopts a MaskRcnn example segmentation algorithm.
7. The implementation method of the deep learning based medicine identification system according to claim 2, wherein: the artificial intelligence algorithm in the S4 comprises a model training algorithm and a recognition algorithm;
the model training algorithm comprises:
a1: obtaining medicine pictures in the medicine blue in an actual scene through a camera, and arranging to obtain a picture library;
a2: marking the medicines in the medicine picture library one by one, wherein the medicines comprise the position, the type, the medicines and the like of each medicine, and dividing a data set into a training set, a verification set and a test set;
a3: establishing a deep learning algorithm model required by a medicine identification task through a research problem;
a4: training an algorithm model by using a training set and a verification set, and iteratively updating parameters at each time until the model converges;
a5: further testing the performance of the model by using the test set, and stopping training if the performance meets the requirements; if the model performance does not meet the requirements, repeating the steps by increasing the data volume and modifying the deep learning algorithm model until the model performance meets the requirements;
the recognition algorithm comprises:
b1 picture acquisition module: shooting a medicine picture by using a camera, and sending the medicine picture to a detection segmentation module;
b2 detection segmentation module: calculating the accurate position of each medicine by using an artificial intelligence algorithm, and dividing the medicines one by one;
b3 identifies the module: the information such as the shape, characters, trademarks, pictures and the like of the medicine box is identified by an artificial intelligence algorithm, and the types of the medicines and the number of each type are accurately identified;
b4 alignment module: comparing and checking the output medicine information with the medicine list of the patient, and if the quantity and the type can be matched, passing; if the comparison is not successful, the pictures, names and number of the medicines which are not successfully matched are measured and output.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811628692.9A CN111382622A (en) | 2018-12-28 | 2018-12-28 | Medicine identification system based on deep learning and implementation method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811628692.9A CN111382622A (en) | 2018-12-28 | 2018-12-28 | Medicine identification system based on deep learning and implementation method thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111382622A true CN111382622A (en) | 2020-07-07 |
Family
ID=71215567
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811628692.9A Pending CN111382622A (en) | 2018-12-28 | 2018-12-28 | Medicine identification system based on deep learning and implementation method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111382622A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112132061A (en) * | 2020-09-24 | 2020-12-25 | 南京中爱人工智能与生命科学研究院有限公司 | Medicine screening method and system based on online identification system |
| CN112347866A (en) * | 2020-10-22 | 2021-02-09 | 上海铂端科技有限公司 | System and method for realizing attachment assembly fool-proof detection processing based on machine vision |
| CN113920335A (en) * | 2021-09-28 | 2022-01-11 | 苏州冷王网络科技有限公司 | Deep learning-based image and text embedded drug label identification method |
| CN114789870A (en) * | 2022-05-20 | 2022-07-26 | 深圳市信成医疗科技有限公司 | Innovative modular drug storage management implementation mode |
| TWI781856B (en) * | 2021-12-16 | 2022-10-21 | 新加坡商鴻運科股份有限公司 | Method for identifying medicine image, computer device and storage medium |
| JP2023005928A (en) * | 2021-06-30 | 2023-01-18 | 株式会社コンテック | Medicine identification system and method for identifying medicine |
| CN116168410A (en) * | 2023-04-21 | 2023-05-26 | 江苏羲辕健康科技有限公司 | Medicine box information identification method and system based on neural network |
-
2018
- 2018-12-28 CN CN201811628692.9A patent/CN111382622A/en active Pending
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112132061A (en) * | 2020-09-24 | 2020-12-25 | 南京中爱人工智能与生命科学研究院有限公司 | Medicine screening method and system based on online identification system |
| CN112347866A (en) * | 2020-10-22 | 2021-02-09 | 上海铂端科技有限公司 | System and method for realizing attachment assembly fool-proof detection processing based on machine vision |
| JP2023005928A (en) * | 2021-06-30 | 2023-01-18 | 株式会社コンテック | Medicine identification system and method for identifying medicine |
| JP7274529B2 (en) | 2021-06-30 | 2023-05-16 | 株式会社コンテック | Drug identification system and drug identification method |
| CN113920335A (en) * | 2021-09-28 | 2022-01-11 | 苏州冷王网络科技有限公司 | Deep learning-based image and text embedded drug label identification method |
| TWI781856B (en) * | 2021-12-16 | 2022-10-21 | 新加坡商鴻運科股份有限公司 | Method for identifying medicine image, computer device and storage medium |
| CN114789870A (en) * | 2022-05-20 | 2022-07-26 | 深圳市信成医疗科技有限公司 | Innovative modular drug storage management implementation mode |
| CN116168410A (en) * | 2023-04-21 | 2023-05-26 | 江苏羲辕健康科技有限公司 | Medicine box information identification method and system based on neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111382622A (en) | Medicine identification system based on deep learning and implementation method thereof | |
| CN109616173B (en) | Medicine management method, medicine management device, computer equipment and storage medium | |
| CN113705733A (en) | Medical bill image processing method and device, electronic device and storage medium | |
| CN112989087B (en) | Image processing method, device and computer readable storage medium | |
| CN112509661B (en) | Methods, computing devices, and media for identifying physical examination reports | |
| KR102122561B1 (en) | Method for recognizing characters on document images | |
| US20240304014A1 (en) | Computer-implemented segmented numeral character recognition and reader | |
| CN110689046A (en) | Image recognition method, image recognition device, computer device, and storage medium | |
| CN107403179B (en) | Registration method and device for article packaging information | |
| CN109215131A (en) | The driving method and device of conjecture face | |
| CN110781709A (en) | Method and system for classifying and identifying drugs via their packaging and/or labeling | |
| US11393597B1 (en) | Prescription label scanner | |
| Yue et al. | Real-time medicine packet recognition system in dispensing medicines for the elderly | |
| CN110209860A (en) | A kind of interpretable garment coordination method and device based on clothes attribute of template-directed | |
| CN116797864B (en) | Auxiliary cosmetic method, device, equipment and storage medium based on intelligent mirror | |
| US20230394587A1 (en) | Systems and Methods for Automatic Recording of Insurance Card Information | |
| CN112418025A (en) | Weight detection method and device based on deep learning | |
| CN202795378U (en) | Medicine rechecking device | |
| CN115063784A (en) | Bill image information extraction method and device, storage medium and electronic equipment | |
| CN111753618A (en) | Image recognition method and device, computer equipment and computer readable storage medium | |
| CN113901934A (en) | Intelligent visual detection method, system and device for large infusion package product | |
| CN112132061A (en) | Medicine screening method and system based on online identification system | |
| CN113782146A (en) | General medicine recommendation method, device, equipment and medium based on artificial intelligence | |
| CN207249688U (en) | A kind of computer marking device based on artificial intelligence character recognition technology | |
| CN208969723U (en) | A kind of face identification payment devices |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20200707 |