CN113205047B - Medicine name identification method, device, computer equipment and storage medium - Google Patents
Medicine name identification method, device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN113205047B CN113205047B CN202110486506.8A CN202110486506A CN113205047B CN 113205047 B CN113205047 B CN 113205047B CN 202110486506 A CN202110486506 A CN 202110486506A CN 113205047 B CN113205047 B CN 113205047B
- Authority
- CN
- China
- Prior art keywords
- image
- medicine
- medicine name
- text
- name
- 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.)
- Active
Links
- 239000003814 drug Substances 0.000 title claims abstract description 566
- 238000000034 method Methods 0.000 title claims abstract description 51
- 229940079593 drug Drugs 0.000 title claims description 84
- 238000001514 detection method Methods 0.000 claims description 43
- 238000000605 extraction Methods 0.000 claims description 24
- 238000004422 calculation algorithm Methods 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 13
- 230000015572 biosynthetic process Effects 0.000 claims description 8
- 238000003786 synthesis reaction Methods 0.000 claims description 8
- 238000005516 engineering process Methods 0.000 abstract description 3
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 16
- 238000013528 artificial neural network Methods 0.000 description 10
- 238000013527 convolutional neural network Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 7
- 238000004364 calculation method Methods 0.000 description 6
- 125000004122 cyclic group Chemical group 0.000 description 5
- 239000002245 particle Substances 0.000 description 4
- 238000011176 pooling Methods 0.000 description 4
- 238000012015 optical character recognition Methods 0.000 description 3
- 230000004438 eyesight Effects 0.000 description 2
- 238000003709 image segmentation Methods 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000003416 augmentation Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/40—Document-oriented image-based pattern recognition
- G06V30/41—Analysis of document content
- G06V30/412—Layout analysis of documents structured with printed lines or input boxes, e.g. business forms or tables
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Evolutionary Biology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The application relates to the field of artificial intelligence, realizes training of a medicine name recognition model according to training images and expansion images and medicine name fuzzy matching of medicine name recognition results output by the medicine name recognition model, and effectively improves the accuracy of recognizing medicine names. To a method, apparatus, computer device and storage medium for identifying medicine names, the method comprising: inputting the image training data and the image expansion data into a medicine name recognition model for iterative training to obtain a trained medicine name recognition model; acquiring a medicine box image to be subjected to medicine name recognition, and determining at least one image to be recognized corresponding to the medicine box image; inputting each image to be identified into a trained medicine name identification model to carry out medicine name identification, and obtaining a medicine name identification result corresponding to the medicine box image; and carrying out medicine name fuzzy matching on the medicine name recognition result based on a preset medicine name information base to obtain a medicine name corresponding to the medicine box image. In addition, the application also relates to a blockchain technology, and a medicine name information base can be stored in the blockchain.
Description
Technical Field
The present application relates to the field of artificial intelligence, and in particular, to a method, an apparatus, a computer device, and a storage medium for identifying a medicine name.
Background
In daily life, the middle-aged and elderly people often contact various medicines, but the characters on many medicine boxes are smaller, and the names of the medicines on the medicine boxes are difficult to identify for the middle-aged and elderly people. The existing method mainly adopts OCR (Optical Character Recognition ) to automatically recognize characters on a medicine box, but the OCR technology is only suitable for simple scenes, and mainly adopts a feature extraction and template matching method in a text recognition stage, utilizes experience knowledge of people to guide character feature extraction, and then matches characters from a feature library according to character similarity. The method has poor stability and effectiveness, and is difficult to identify the complex and diverse drug names of the drug box.
It is therefore a great need to address how to improve the accuracy of identifying the drug name on a drug cassette.
Disclosure of Invention
The application provides a medicine name recognition method, a medicine name recognition device, computer equipment and a storage medium, wherein the medicine name recognition model is trained according to image training data and image expansion data, so that the trained medicine name recognition model is suitable for recognizing scenes of complex and various medicine names in a medicine box image, and the accuracy of recognizing the medicine names can be effectively improved; the accuracy of identifying the medicine name is further improved by carrying out medicine name fuzzy matching on the medicine name identification result output by the medicine name identification model.
In a first aspect, the present application provides a method for identifying a drug name, the method comprising:
acquiring image training data, wherein the image training data is obtained by extracting text from a sample medicine box image and acquiring image expansion data;
Inputting the image training data and the image expansion data into a medicine name recognition model for iterative training until the medicine name recognition model converges, so as to obtain the trained medicine name recognition model;
Acquiring a medicine box image to be subjected to medicine name recognition, and determining at least one image to be recognized corresponding to the medicine box image;
Inputting each image to be identified into the trained medicine name identification model to carry out medicine name identification, and obtaining a medicine name identification result corresponding to the medicine box image;
And carrying out medicine name fuzzy matching on the medicine name recognition result based on a preset medicine name information base to obtain the medicine name corresponding to the medicine box image.
In a second aspect, the present application also provides a medicine name recognition device, the device comprising:
The image data acquisition module is used for acquiring image training data, wherein the image training data is obtained by extracting text from a sample medicine box image and acquiring image expansion data;
The model training module is used for inputting the image training data and the image expansion data into a medicine name recognition model for iterative training until the medicine name recognition model converges to obtain the trained medicine name recognition model;
The image extraction module is used for acquiring a medicine box image to be subjected to medicine name recognition and determining at least one image to be recognized corresponding to the medicine box image;
The medicine name recognition module is used for inputting each image to be recognized into the trained medicine name recognition model to carry out medicine name recognition, and obtaining a medicine name recognition result corresponding to the medicine box image;
and the medicine name fuzzy matching module is used for carrying out medicine name fuzzy matching on the medicine name recognition result based on a preset medicine name information base to obtain the medicine name corresponding to the medicine box image.
In a third aspect, the present application also provides a computer device comprising a memory and a processor;
The memory is used for storing a computer program;
The processor is configured to execute the computer program and implement the medicine name identification method as described above when the computer program is executed.
In a fourth aspect, the present application also provides a computer readable storage medium storing a computer program which, when executed by a processor, causes the processor to implement a method for identifying a drug name as described above.
The application discloses a medicine name recognition method, a medicine name recognition device, computer equipment and a storage medium, wherein the image training data are acquired, the image expansion data are acquired, and the image training data and the image expansion data are input into a medicine name recognition model for iterative training, so that the trained medicine name recognition model is suitable for recognizing the scenes of complex and various medicine names in a medicine box image, and the accuracy of recognizing the medicine names can be effectively improved; the method comprises the steps of obtaining a medicine box image to be subjected to medicine name recognition and determining the image to be recognized, so that the image to be recognized of a medicine box text can be obtained; by inputting each image to be identified into a trained medicine name identification model for medicine name identification, the accuracy of medicine name identification results can be effectively improved; and (3) carrying out medicine name fuzzy matching on medicine name recognition results based on a preset medicine name information base to obtain medicine names corresponding to the medicine box images, so that the accuracy of medicine name recognition on the medicine box images is further improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present application, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic flow chart of a medicine name identification method provided by an embodiment of the application;
FIG. 2 is a schematic diagram for training a drug name recognition model according to an embodiment of the present application;
FIG. 3 is a schematic flow chart of a sub-step of acquiring image expansion data provided by an embodiment of the present application;
FIG. 4 is a schematic flow chart of a sub-step of determining an image to be identified provided by an embodiment of the present application;
FIG. 5 is a schematic flow chart of a substep of text detection of a drug cassette image provided by an embodiment of the present application;
FIG. 6 is a schematic block diagram of a drug name recognition device provided by an embodiment of the present application;
fig. 7 is a schematic block diagram of a computer device according to an embodiment of the present application.
Detailed Description
The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
The flow diagrams depicted in the figures are merely illustrative and not necessarily all of the elements and operations/steps are included or performed in the order described. For example, some operations/steps may be further divided, combined, or partially combined, so that the order of actual execution may be changed according to actual situations.
It is to be understood that the terminology used in the description of the application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used in this specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
It should also be understood that the term "and/or" as used in the present specification and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes such combinations.
The embodiment of the application provides a medicine name identification method, a medicine name identification device, computer equipment and a storage medium. The medicine name recognition method can be applied to a server or a terminal, and the medicine name recognition model is trained according to the training image and the expansion image, so that the trained medicine name recognition model is suitable for recognizing the scenes of complex and various medicine names in the medicine box image, and the accuracy of recognizing the medicine names can be effectively improved; the accuracy of identifying the medicine name is further improved by carrying out medicine name fuzzy matching on the medicine name identification result output by the medicine name identification model.
The servers may be independent servers or may be server clusters. The terminal can be electronic equipment such as a smart phone, a tablet computer, a notebook computer, a desktop computer and the like.
Some embodiments of the present application are described in detail below with reference to the accompanying drawings. The following embodiments and features of the embodiments may be combined with each other without conflict.
As shown in fig. 1, the medicine name recognition method includes steps S10 to S50.
Step S10, acquiring image training data, wherein the image training data is obtained by extracting text from a sample medicine box image, and acquiring image expansion data.
In the embodiment of the application, in order to improve the accuracy of the medicine name recognition model in recognizing the medicine name, the initial medicine name recognition model needs to be trained until convergence before the medicine name recognition model is adopted to recognize the medicine name of the image to be recognized, so that a trained medicine name recognition model is obtained.
Referring to fig. 2, fig. 2 is a schematic diagram illustrating training of a drug name recognition model according to an embodiment of the present application. As shown in fig. 2, a specific training process may include: acquiring image training data, wherein the image training data is obtained by extracting text from a sample medicine box image, and acquiring image expansion data; and inputting the image training data and the image expansion data into an initial medicine name recognition model for iterative training until the medicine name recognition model converges, and obtaining a trained medicine name recognition model.
For example, text extraction may be performed on a predetermined number of sample cartridge images to obtain image training data. The text detection model is used for extracting the text of the sample medicine box image.
For example, a preset number of image augmentation data may be generated. It should be noted that, since the characters on the medicine box image are generally complex and include characters with different font types, different sizes and different directions, the medicine name recognition model needs to be trained by adding image expansion data, so that the trained medicine name recognition model is suitable for recognizing the scenes of complex and various medicine names in the medicine box image, and the accuracy of the medicine name recognition model is improved.
Referring to fig. 3, fig. 3 is a schematic flowchart of a sub-step of acquiring image expansion data according to an embodiment of the present application, which may specifically include the following steps S101 to S103.
Step S101, based on a preset medicine knowledge base, acquiring text data containing medicine information, wherein the text data comprises at least one of Chinese, english, numbers and symbols.
For example, a predetermined amount of text data containing medication information may be extracted from a medication knowledge base. Wherein the extracted text data is plain text; the content of the text data may include at least one of chinese, english, numerals, and symbols.
And S102, adding text attributes to the text data to obtain the text data with the text attributes added, wherein the text attributes comprise at least one of text length, text direction, font size and font style.
For example, different text attributes may be added to the text data. For example, the text direction of the text data is set to the landscape direction; for example, the font size of the text data is set to No. three; for another example, the font style of the text data is set to color. So that a plurality of text data with text attributes added thereto can be obtained.
For example, the text direction of the text data after adding the text attribute may be a lateral direction and the font style may be bolded.
For example, the text direction of the text data after adding the text attribute may be portrait, and the font size may be No. three.
And step S103, adding the text data with the text attribute added into a preset image template to obtain the image expansion data.
It should be noted that the preset image template may be a blank image.
For example, text data to which text attributes are added may be loaded into a blank image, thereby obtaining image extension data.
By constructing the image expansion data, the method can automatically generate rich image training data, solves the problem of insufficient training samples, and improves the robustness and accuracy of the medicine name recognition model.
And step S20, inputting the image training data and the image expansion data into a medicine name recognition model for iterative training until the medicine name recognition model converges, and obtaining the trained medicine name recognition model.
In the embodiment of the application, the medicine name recognition model is trained by inputting the added image expansion data and the image training data together, so that the trained medicine name recognition model is suitable for recognizing the scenes of complex and various medicine names in the medicine box image, thereby improving the accuracy of the medicine name recognition model.
The medicine name recognition model is used for recognizing text information in the image, classifying the text information by labels, and determining the label with the highest output probability as a medicine name recognition result. The drug name recognition model may include a convolutional neural network and a recurrent neural network; the convolution neural network is used for convoluting and pooling the input image and outputting a characteristic image; the cyclic neural network is used for carrying out medicine name classification prediction on the characteristic images and outputting medicine name recognition results.
Illustratively, when training the drug name recognition model, the loss function value of the model during training can be calculated through a CTC loss function, and parameters of the model are adjusted according to the loss function value. The CTC (Connectionist Temporal Classification, connection timing classification) loss function is used to solve the alignment problem of the input features and the output tags.
In some embodiments, inputting the image training data and the image expansion data into the drug name recognition model for iterative training until the drug name recognition model converges may include: determining training sample images of each training round according to the image training data and the image expansion data; inputting the current training sample image into a convolutional neural network for convolution and pooling, and outputting a characteristic image; inputting the characteristic images into a cyclic neural network to conduct medicine name classification prediction, and outputting corresponding medicine name recognition results; classifying the loss function based on the connection time sequence, and determining a loss function value corresponding to the medicine name recognition result; and if the loss function value is greater than a preset loss value threshold, adjusting parameters of the convolutional neural network and the cyclic neural network, and performing the next training until the obtained loss function value is less than or equal to the loss value threshold, ending the training, and obtaining a trained medicine name recognition model.
For example, the preset loss value threshold may be set according to practical situations, and the specific value is not limited herein.
For example, the parameters of the convolutional neural network and the recurrent neural network may be adjusted based on an error back-propagation algorithm, or may be adjusted according to other algorithms, such as a gradient descent algorithm. Wherein the error back propagation (Error Back Propagation, BP) algorithm is a multi-layer feedforward neural network trained in accordance with the error back propagation algorithm.
In order to further ensure the privacy and security of the trained drug name recognition model, the trained drug name recognition model may also be stored in a node of a blockchain. When a trained drug name recognition model is required, it can be invoked from the nodes of the blockchain.
By classifying the loss function based on the connection time sequence, calculating the loss function value and adjusting the parameters of the convolutional neural network and the cyclic neural network according to the loss function value, the alignment problem of the input characteristics and the output labels is solved, and the accuracy of identifying the medicine name by the trained medicine name identification model is effectively improved.
In some embodiments, the method for inputting the current training sample image into the convolutional neural network for convolution and pooling, and before outputting the characteristic image, the method may further include: determining an image recognition direction corresponding to the medicine name recognition model; determining the image direction corresponding to the current wheel training sample image according to the height and the width corresponding to the current wheel training sample image; performing direction adjustment on the training sample images with different image directions and image recognition directions to obtain adjusted training sample images; and inputting the adjusted training sample image and the training sample image which does not need to be subjected to direction adjustment into a convolutional neural network to carry out convolution and pooling, so as to obtain a characteristic image corresponding to the current training sample image.
When the medicine name recognition model recognizes an image, the same medicine name recognition model is applicable to recognizing an image in a fixed direction in order to ensure the accuracy of recognition. If training sample images in multiple directions exist, multiple medicine name recognition models need to be trained for recognition. Therefore, when training the medicine name recognition model, it is necessary to determine the image recognition direction of the medicine name recognition model first and then perform direction adjustment on the training sample image.
For example, an image recognition direction corresponding to the drug name recognition model may be set; for example, the image recognition direction may be set to be horizontal or the image recognition direction may be set to be vertical.
For example, the image direction corresponding to the current wheel training sample image may be determined according to the height and width corresponding to the current wheel training sample image. For example, when the width of the training sample image is greater than the height, the image direction of the training sample image may be determined to be lateral; when the width of the training sample image is less than or equal to the height, the image direction of the training sample image may be determined to be longitudinal.
For example, if the image recognition direction of the drug name recognition model is horizontal, the training sample image whose image direction is vertical needs to be subjected to direction adjustment. For example, the training sample image may be rotated 90 °. For training sample images of other image directions, rotation may be performed according to the actual image direction.
By unifying the directions of the training sample images, the training of the transverse images and the longitudinal images by one medicine name recognition model can be realized, the training of the transverse images by one medicine name recognition model is not needed, the training of the longitudinal images by the other medicine name recognition model is not needed, and the training process is simplified.
Step S30, acquiring a medicine box image to be subjected to medicine name recognition, and determining at least one image to be recognized corresponding to the medicine box image.
It should be noted that, the medicine name identification method provided by the embodiment of the application can be applied to a complex scene for identifying complex and diverse medicine names in a medicine box image. The user can upload the medicine box image which needs medicine name identification to a server or a terminal, wherein the server or the terminal is provided with a medicine name identification system or a medicine name identification application program. For example, a user may take a medicine box image requiring medicine name identification through a mobile phone, and then upload a medicine name identification system or a medicine name identification application program to perform medicine name identification.
When an image uploading operation of a user is detected, the medicine box image uploaded by the user is determined to be the medicine box image to be subjected to medicine name identification according to the image uploading operation.
In the embodiment of the application, after the medicine box image to be subjected to medicine name recognition is acquired, at least one image to be recognized corresponding to the medicine box image can be determined. Referring to fig. 4, fig. 4 is a schematic flowchart of a sub-step of determining an image to be identified according to an embodiment of the present application, which may specifically include the following step S301 and step S302.
Step S301, inputting the medicine box image into a text detection model for text detection, and obtaining text position information corresponding to the medicine box image.
It should be noted that the text detection model may be a DB-Net (Differentiable Binarization Network, differential binarizable network) model. The biggest innovation point of the DB-Net model is that each pixel point is subjected to self-adaptive binarization, wherein a binarization threshold value is obtained through network learning, and the process of binarization is added into a network to be trained together, so that the robustness of the text detection model can be effectively enhanced, and the detection speed of the text detection model can be improved. In addition, a backbone network of the DB-Net model adopts ResNet (Residual Neural Network ) structure, and in the training process, after the operations of feature extraction, upsampling fusion and the like are carried out after the picture is input, a probability feature map and a threshold feature map are respectively predicted according to the obtained feature images, and a binarization feature map is obtained through calculation of the probability feature map and the threshold feature map.
In the embodiment of the application, the text detection model at least comprises a feature extraction layer; the feature extraction layer is used for extracting features of the images input into the text detection model to obtain corresponding feature images. The text detection model may include a feature prediction layer, a binarization layer, and a text position prediction layer in addition to the feature extraction layer. The feature prediction layer is used for predicting a probability feature image and a threshold feature image corresponding to the feature image; the binarization layer is used for performing binarization calculation on the probability feature image and the threshold feature image and outputting a binarization feature image; the text position prediction layer is used for carrying out text region identification on the binarized characteristic image and outputting text position information.
By way of example, the text detection model may be a pre-trained text detection model. In some embodiments, a preset number of sample cartridge images may be acquired, and the sample cartridge images may be input into an initial text detection model for iterative training until the text detection model converges, to obtain a trained text detection model. The specific training process is not limited herein.
In some embodiments, the medicine box image to be subjected to medicine name recognition can be input into a trained text detection model to perform text detection, and text position information corresponding to the medicine box image is obtained.
Referring to fig. 5, fig. 5 is a schematic flowchart of a substep of text detection on a medicine box image according to an embodiment of the present application, which may specifically include the following steps S3011 to S3013.
Step S3011, inputting the medicine box image into the feature extraction layer for feature extraction, and obtaining a feature image corresponding to the medicine box image.
For example, the medicine box image may be input into the feature extraction layer, the feature extraction layer performs feature extraction, and the feature image corresponding to the medicine box image is output. The feature extraction layer may include, among other things, a FPN (Feature Pyramid Networks, feature pyramid network). It should be noted that the FPN solves the multi-scale problem in object detection, and greatly improves the performance of small object detection under the condition of not increasing the calculation amount of the original model basically by simple network connection change.
For example, the image of the medicine box can be up-sampled through the FPN network, so that the characteristic image corresponding to the image of the medicine box is extracted.
Step S3012, determining a binarization feature map corresponding to the feature image.
For example, the binarized feature map corresponding to the feature image is determined, the feature image output by the feature extraction layer may be input into the feature prediction layer, and the probability feature map corresponding to the feature image and the threshold feature image may be obtained by prediction by the feature prediction layer. And then inputting the probability feature image and the threshold feature image into a binarization layer for binarization calculation, and outputting a binarization feature image.
Illustratively, the binarization layer may perform binarization calculation on the probability feature map and the threshold feature image based on a micro-binarizable formula to obtain a binarized feature map.
Step S3013, determining a text region in the binarized feature map, and determining the text position information according to the text region.
For example, in determining the text region in the binarized feature map, the text region in the binarized feature map may be determined based on pixel values in the binarized feature map. For example, a pixel point whose pixel value is greater than a preset pixel threshold value is determined as a text region. And determining the pixel point with the pixel value smaller than the preset pixel threshold value as a non-text area. The preset pixel threshold value may be set according to practical situations, and specific numerical values are not limited herein.
For example, the upper left corner coordinates and the lower right corner coordinates of the text region may be determined as text position information. So that the text position information may include the upper left and lower right coordinates of the text region.
In the conventional text detection method based on image segmentation, a fixed threshold is usually set to convert a probability feature map generated by a segmentation network into a binarization feature map. Because different thresholds have a larger influence on the performance of the model, the conventional text detection method based on image segmentation has lower accuracy. The DB-Net model in the embodiment of the application can adaptively predict the threshold value of each position in the image by inserting the binarization operation into the segmentation network for joint optimization, thereby being capable of completely distinguishing the foreground and the background of the pixel. Not only the detection speed of the text detection model is improved, but also the robustness of the text detection model is enhanced.
By inputting the medicine box image to be subjected to medicine name recognition into the text detection model for text detection, text position information corresponding to texts in different formats in the medicine box image can be accurately determined.
And step S302, segmenting the medicine box image according to the text position information to obtain at least one image to be identified.
For example, an image area to be segmented can be determined according to the upper left corner coordinate and the lower right corner coordinate of the text area, and then screenshot is performed on the image area to obtain an image to be identified. Of course, other splitting methods are also possible, and the specific splitting method is not limited herein. And correspondingly segmenting each piece of text position information to obtain an image to be identified.
The medicine box image is segmented according to the text position information, so that the image to be identified containing the text can be accurately obtained, and the accuracy of medicine name identification of the image to be identified can be improved.
And S40, inputting each image to be identified into a trained medicine name identification model to carry out medicine name identification, and obtaining a medicine name identification result corresponding to the medicine box image.
After at least one to-be-identified image corresponding to the medicine box image to be subjected to medicine name identification is determined, each to-be-identified image can be input into a trained medicine name identification model to carry out medicine name identification, and therefore medicine name identification results corresponding to the medicine box image can be obtained.
In some embodiments, before inputting each image to be identified into the medicine name identification model for medicine name identification, the method may further include: determining an image recognition direction corresponding to the medicine name recognition model; determining an image to be identified for direction adjustment and an image to be identified without direction adjustment according to the image identification direction; and carrying out direction adjustment on the image to be identified to be subjected to direction adjustment to obtain an adjusted image to be identified.
The image recognition direction corresponding to the medicine name recognition model is determined during training. For example, if the image recognition direction to which the medicine name recognition model corresponds during training is a lateral direction, it may be determined that the image recognition direction to which the medicine name recognition model corresponds at this time is a lateral direction. For example, if the image recognition direction corresponding to the medicine name recognition model is vertical during training, it is possible to determine that the image recognition direction corresponding to the medicine name recognition model at this time is vertical.
For example, after determining the image recognition direction corresponding to the drug name recognition model, the image to be recognized to be subjected to direction adjustment and the image to be recognized not requiring direction adjustment may be determined according to the image recognition direction. For example, when the image recognition direction is transverse, if the width of the image to be recognized is greater than the height, which indicates that the image direction of the image to be recognized is transverse, it may be determined that the image to be recognized does not need direction adjustment. For example, when the image recognition direction is the horizontal direction, if the width of the image to be recognized is smaller than the height, which indicates that the image direction of the image to be recognized is the vertical direction, it may be determined that the image to be recognized needs to be subjected to direction adjustment.
For example, the image to be recognized, which is to be subjected to the direction adjustment, may be rotated clockwise or counterclockwise until the image direction of the image to be recognized is adjusted to be a lateral direction.
In some embodiments, inputting each image to be identified into the drug name identification model for drug name identification may include: and inputting the adjusted image to be identified and the image to be identified which does not need direction adjustment into a medicine name identification model to carry out medicine name identification.
For example, the adjusted image to be identified and the image to be identified which does not need direction adjustment can be input into a medicine name identification model to carry out medicine name identification, so that a medicine name identification result corresponding to the medicine box image can be obtained.
By determining the image recognition direction corresponding to the medicine name recognition model, whether the direction of the image to be recognized needs to be adjusted can be judged, so that one medicine name recognition model can recognize the transverse and longitudinal image to be recognized, and the other medicine name recognition model does not need to be additionally added for recognition, so that the calculation amount and the memory consumption are reduced, and the accuracy of medicine name recognition by the medicine name recognition model is improved.
And step S50, based on a preset medicine name information base, performing medicine name fuzzy matching on the medicine name recognition result to obtain a medicine name corresponding to the medicine box image.
It should be noted that, because the medicine name recognition result obtained by the medicine name recognition model may have homonyms or synonyms, the recognition result is not accurate enough, and therefore, by performing medicine name fuzzy matching on the medicine name recognition result, a more accurate medicine name can be obtained.
Illustratively, the drug name information library includes a variety of standard drug name texts. In the embodiment of the application, various standard medicine name texts can be collected in advance and stored in a medicine name information base. To further ensure the privacy and security of the drug name information library, the drug name information library may be stored in a node of a blockchain.
In some embodiments, based on a preset medicine name information base, performing medicine name fuzzy matching on the medicine name recognition result to obtain a medicine name corresponding to the medicine box image may include: determining an edit distance value between a medicine name recognition result and a standard medicine name text in a medicine name information base based on a preset edit distance algorithm; and determining the standard medicine name text corresponding to which the edit distance value is smaller than the preset edit distance threshold value as the medicine name corresponding to the medicine box image.
Note that LEVENSHTEIN DISTANCE (edit distance) algorithm refers to the minimum number of editing operations required to change from one to another between two strings.
For example, an edit distance value between the drug name recognition result and a standard drug name text in the drug name information base can be calculated based on an edit distance algorithm, and a drug name corresponding to a drug box image is determined by determining a standard drug name text corresponding to an edit distance value smaller than a preset edit distance threshold.
The preset edit distance threshold may be set according to practical situations, and specific numerical values are not limited herein.
By performing medicine name fuzzy matching on the medicine name recognition result based on a preset medicine name information base, the accuracy of the recognized medicine name can be further improved.
In some embodiments, when identifying complex and diverse drug names in a drug cassette image, for example, identifying drug names, such as drug names "AB-C particles", which are composed of chinese, english and symbols, have a lateral text direction and a color font style, the drug cassette image is first input into a text detection model to perform text detection, and text position information corresponding to the drug cassette image is obtained. Secondly, segmenting the medicine box image according to the text position information to obtain at least one image to be identified, wherein the image to be identified can be an image containing a medicine name or an image containing other characters. And thirdly, carrying out direction adjustment on the image to be identified to be subjected to direction adjustment according to the image identification direction of the medicine name identification model, inputting the adjusted image to be identified and the image to be identified which does not need direction adjustment into the medicine name identification model to carry out medicine name identification, and obtaining medicine name identification results corresponding to the medicine box image, such as 'AB_D particles'. And finally, carrying out medicine name fuzzy matching on the medicine name identification result 'AB_D particles' based on a preset medicine name information base to obtain medicine name 'AB-C particles' corresponding to the medicine box image. So that the drug name in the drug cassette image can be accurately identified.
In some embodiments, after obtaining the drug name corresponding to the drug cassette image, it may further include: and determining the medication instruction information of the medicine name corresponding to the medicine box image based on the corresponding relation between the preset medicine name and the medication instruction information, and displaying the medicine name and the medication instruction information.
For example, the medication instruction information may include indication function information, administration instruction information, and the like.
For example, the medication name and medication instruction information may be displayed in a server or terminal where the user uploads the medication box image. For example, the medication name and medication instruction information are displayed on the user's mobile phone.
In some embodiments, after obtaining the drug name corresponding to the drug cassette image, it may further include: and performing voice synthesis on the medicine names corresponding to the medicine box images, generating voice information corresponding to the medicine names, and broadcasting the voice information.
For example, the voice information corresponding To the drug name may be generated by voice synthesis of the drug name corresponding To the drug cassette image through a TTS (Text To Speech) model. It should be noted that, the TTS model includes modules such as text analysis, acoustic model, and audio synthesis, and may implement conversion of a text segment into a speech signal.
For example, after the voice information corresponding to the pharmaceutical name is generated, the voice information may be broadcasted on a server or a terminal. For example, voice information is broadcast by a mobile handset of the user. In addition, the medication instruction information can be synthesized by voice and broadcast.
In some embodiments, after obtaining the drug name corresponding to the drug cassette image, it may further include: determining the medicine instruction information of the medicine name corresponding to the medicine box image based on the corresponding relation between the preset medicine name and the medicine instruction information, and displaying the medicine name and the medicine instruction information; and performing voice synthesis on the medicine names corresponding to the medicine box images, generating voice information corresponding to the medicine names, and broadcasting the voice information.
The medicine name and the medicine instruction information on the medicine box image are more conveniently acquired by a user through determining the medicine instruction information of the medicine name corresponding to the medicine box image and displaying the medicine name and the medicine instruction information, so that the medicine box image is more humanized; through carrying out voice synthesis to the medicine names corresponding to the medicine box images, generating voice information corresponding to the medicine names and broadcasting the voice information, the medicine names in the medicine box images are also very conveniently acquired by users with poor eyesight, and the experience of the users is improved.
According to the medicine name recognition method provided by the embodiment, the image expansion data are constructed, so that abundant image training data can be automatically generated, the problem of insufficient training samples is solved, and the robustness and accuracy of a medicine name recognition model are improved; the added image expansion data and the image training data are input into the medicine name recognition model together for training, so that the trained medicine name recognition model is suitable for recognizing scenes of complex and various medicine names in the medicine box image, and the accuracy of the medicine name recognition model is improved; by classifying the loss function based on the connection time sequence, calculating the loss function value and adjusting the parameters of the convolutional neural network and the cyclic neural network according to the loss function value, the alignment problem of the input characteristics and the output labels is solved, and the accuracy of identifying the medicine name by the trained medicine name identification model is effectively improved; by unifying the directions of the training sample images, the training of the transverse images and the longitudinal images by one medicine name recognition model can be realized without independently training the transverse images of one medicine name recognition model and independently training the longitudinal images of the other medicine name recognition model, so that the training process is simplified; the medicine box image to be subjected to medicine name recognition is input into a text detection model to carry out text detection, so that text position information corresponding to texts in different formats in the medicine box image can be accurately determined; the medicine box image is segmented according to the text position information, so that an image to be identified containing the text can be accurately obtained, and the accuracy of medicine name identification of the image to be identified can be improved; the accuracy of the identified medicine names can be further improved by carrying out medicine name fuzzy matching on medicine name identification results based on a preset medicine name information base; the medicine name and the medicine instruction information on the medicine box image are more conveniently acquired by a user through determining the medicine instruction information of the medicine name corresponding to the medicine box image and displaying the medicine name and the medicine instruction information, so that the medicine box image is more humanized; through carrying out voice synthesis to the medicine names corresponding to the medicine box images, generating voice information corresponding to the medicine names and broadcasting the voice information, the medicine names in the medicine box images are also very conveniently acquired by users with poor eyesight, and the experience of the users is improved.
Referring to fig. 6, fig. 6 is a schematic block diagram of a medicine name recognition device 1000 according to an embodiment of the present application, where the medicine name recognition device is configured to perform the aforementioned medicine name recognition method. The medicine name recognition device can be configured in a server or a terminal.
As shown in fig. 6, the medicine name recognition device 1000 includes: an image data acquisition module 1001, a model training module 1002, an image extraction module 1003, a medicine name recognition module 1004, and a medicine name fuzzy matching module 1005.
The image data obtaining module 1001 is configured to obtain image training data, where the image training data is obtained by extracting text from a sample medicine box image, and obtain image expansion data.
The model training module 1002 is configured to input the image training data and the image expansion data into a drug name recognition model for iterative training until the drug name recognition model converges, so as to obtain a trained drug name recognition model.
The image extraction module 1003 is configured to acquire a medicine box image to be subjected to medicine name recognition, and determine at least one image to be recognized corresponding to the medicine box image.
And a medicine name recognition module 1004, configured to input each image to be recognized into the trained medicine name recognition model to perform medicine name recognition, and obtain a medicine name recognition result corresponding to the medicine box image.
And a medicine name fuzzy matching module 1005, configured to perform medicine name fuzzy matching on the medicine name recognition result based on a preset medicine name information base, so as to obtain a medicine name corresponding to the medicine box image.
It should be noted that, for convenience and brevity of description, the specific working process of the apparatus and each module described above may refer to the corresponding process in the foregoing method embodiment, which is not described herein again.
The apparatus described above may be implemented in the form of a computer program which is executable on a computer device as shown in fig. 7.
Referring to fig. 7, fig. 7 is a schematic block diagram of a computer device according to an embodiment of the present application. The computer device may be a server or a terminal.
Referring to fig. 7, the computer device includes a processor and a memory connected by a system bus, wherein the memory may include a non-volatile storage medium and an internal memory.
The processor is used to provide computing and control capabilities to support the operation of the entire computer device.
The internal memory provides an environment for the execution of a computer program in a non-volatile storage medium that, when executed by a processor, causes the processor to perform any of a number of drug name recognition methods.
It should be appreciated that the Processor may be a central processing unit (Central Processing Unit, CPU), it may also be other general purpose processors, digital signal processors (DIGITAL SIGNAL Processor, DSP), application SPECIFIC INTEGRATED Circuit (ASIC), field-Programmable gate array (Field-Programmable GATE ARRAY, FPGA) or other Programmable logic device, discrete gate or transistor logic device, discrete hardware components, or the like. Wherein the general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
Wherein in one embodiment the processor is configured to run a computer program stored in the memory to implement the steps of:
Acquiring image training data, wherein the image training data is obtained by extracting text from a sample medicine box image and acquiring image expansion data; inputting the image training data and the image expansion data into a medicine name recognition model for iterative training until the medicine name recognition model converges, so as to obtain the trained medicine name recognition model; acquiring a medicine box image to be subjected to medicine name recognition, and determining at least one image to be recognized corresponding to the medicine box image; inputting each image to be identified into the trained medicine name identification model to carry out medicine name identification, and obtaining a medicine name identification result corresponding to the medicine box image; and carrying out medicine name fuzzy matching on the medicine name recognition result based on a preset medicine name information base to obtain the medicine name corresponding to the medicine box image.
In one embodiment, the processor, when implementing the acquisition of the image expansion data, is configured to implement:
Based on a preset medicine knowledge base, acquiring text data containing medicine information, wherein the text data comprises at least one of Chinese, english, numbers and symbols; adding text attributes to the text data to obtain the text data added with the text attributes, wherein the text attributes comprise at least one of text length, text direction, font size and font style; and adding the text data with the text attribute added into a preset image template to obtain the image expansion data.
In one embodiment, the processor, when implementing determining at least one image to be identified corresponding to the cartridge image, is configured to implement:
Inputting the medicine box image into a text detection model for text detection to obtain text position information corresponding to the medicine box image; and cutting the medicine box image according to the text position information to obtain at least one image to be identified.
In one embodiment, the text detection model includes at least a feature extraction layer; the processor is used for realizing that when realizing that the medicine box image is input into a text detection model to carry out text detection and obtaining text position information corresponding to the medicine box image, the processor is used for realizing that:
Inputting the medicine box image into the feature extraction layer for feature extraction to obtain a feature image corresponding to the medicine box image; determining a binarization feature map corresponding to the feature image; and determining a text region in the binarization feature map, and determining the text position information according to the text region.
In one embodiment, before implementing the drug name recognition by inputting each image to be recognized into the drug name recognition model, the processor is further configured to implement:
Determining an image recognition direction corresponding to the medicine name recognition model; determining an image to be identified to be subjected to direction adjustment and an image to be identified which does not need direction adjustment according to the image identification direction; and carrying out direction adjustment on the image to be identified to be subjected to direction adjustment, and obtaining the adjusted image to be identified.
In one embodiment, the processor is configured to, when implementing drug name recognition by inputting each of the images to be recognized into a drug name recognition model, implement:
and inputting the adjusted image to be identified and the image to be identified which does not need direction adjustment into the medicine name identification model to carry out medicine name identification.
In one embodiment, the drug name information library includes a plurality of standard drug name texts; the processor is used for realizing when realizing that the medicine name is matched with the medicine name identification result based on a preset medicine name information base and obtaining the medicine name corresponding to the medicine box image:
determining an edit distance value between the medicine name recognition result and a standard medicine name text in the medicine name information base based on a preset edit distance algorithm; and determining the standard medicine name text corresponding to the edit distance value smaller than the preset edit distance threshold as the medicine name corresponding to the medicine box image.
In one embodiment, the processor, after implementing obtaining the drug name corresponding to the drug cassette image, is further configured to implement:
determining the medicine instruction information of the medicine name corresponding to the medicine box image based on the corresponding relation between the preset medicine name and the medicine instruction information, and displaying the medicine name and the medicine instruction information; and/or performing voice synthesis on the medicine name corresponding to the medicine box image, generating voice information corresponding to the medicine name, and broadcasting the voice information.
The embodiment of the application also provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, the computer program comprises program instructions, and the processor executes the program instructions to realize any medicine name identification method provided by the embodiment of the application.
The computer readable storage medium may be an internal storage unit of the computer device according to the foregoing embodiment, for example, a hard disk or a memory of the computer device. The computer readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a smart memory Card (SMART MEDIA CARD, SMC), a Secure digital Card (Secure DIGITAL CARD, SD Card), a flash memory Card (FLASH CARD), etc. that are provided on the computer device.
Further, the computer-readable storage medium may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function, and the like; the storage data area may store data created from the use of blockchain nodes, and the like.
The blockchain is a novel application mode of computer technologies such as distributed data storage, point-to-point transmission, consensus mechanism, encryption algorithm and the like. The blockchain (Blockchain), essentially a de-centralized database, is a string of data blocks that are generated in association using cryptographic methods, each of which contains information from a batch of network transactions for verifying the validity (anti-counterfeit) of its information and generating the next block. The blockchain may include a blockchain underlying platform, a platform product services layer, an application services layer, and the like.
While the application has been described with reference to certain preferred embodiments, it will be understood by those skilled in the art that various changes and substitutions of equivalents may be made and equivalents will be apparent to those skilled in the art without departing from the scope of the application. Therefore, the protection scope of the application is subject to the protection scope of the claims.
Claims (8)
1. A method for identifying a drug name, comprising:
Acquiring image training data, wherein the image training data is obtained by extracting texts from sample medicine box images, and acquiring image expansion data, the image expansion data is text data containing medicine information, and the text data is texts with different text lengths, different text directions, different font sizes and different font styles;
Inputting the image training data and the image expansion data into a medicine name recognition model for iterative training until the medicine name recognition model converges, so as to obtain the trained medicine name recognition model;
Acquiring a medicine box image to be subjected to medicine name recognition, and determining at least one image to be recognized corresponding to the medicine box image;
Inputting each image to be identified into the trained medicine name identification model to carry out medicine name identification, and obtaining a medicine name identification result corresponding to the medicine box image;
based on a preset medicine name information base and an edit distance algorithm, performing medicine name fuzzy matching on the medicine name recognition result to obtain a medicine name corresponding to the medicine box image;
After the medicine names corresponding to the medicine box images are obtained, the method further comprises the following steps: based on the corresponding relation between the preset medicine name and the medicine instruction information, determining the medicine instruction information of the medicine name corresponding to the medicine box image, performing voice synthesis on the medicine name corresponding to the medicine box image and the medicine instruction information, generating voice information corresponding to the medicine name and the medicine instruction information, and broadcasting the voice information;
Before each image to be identified is input into the medicine name identification model for medicine name identification, the method further comprises the following steps: determining an image recognition direction corresponding to the medicine name recognition model; determining an image to be identified to be subjected to direction adjustment and an image to be identified which does not need direction adjustment according to the image identification direction; performing direction adjustment on the image to be identified to be subjected to direction adjustment to obtain an adjusted image to be identified;
Inputting each image to be identified into a medicine name identification model for medicine name identification, wherein the method comprises the following steps: inputting the adjusted image to be identified and the image to be identified which does not need direction adjustment into the medicine name identification model for medicine name identification;
The medicine name information base comprises a plurality of standard medicine name texts; the medicine name fuzzy matching is carried out on the medicine name recognition result based on a preset medicine name information base and an edit distance algorithm to obtain a medicine name corresponding to the medicine box image, and the medicine name fuzzy matching comprises the following steps: determining an edit distance value between the medicine name recognition result and a standard medicine name text in the medicine name information base based on the edit distance algorithm; and determining the standard medicine name text corresponding to the edit distance value smaller than the preset edit distance threshold as the medicine name corresponding to the medicine box image.
2. The method of claim 1, wherein the acquiring image extension data comprises:
Based on a preset medicine knowledge base, acquiring text data containing medicine information, wherein the text data comprises at least one of Chinese, english, numbers and symbols;
Adding text attributes to the text data to obtain the text data added with the text attributes, wherein the text attributes comprise at least one of text length, text direction, font size and font style;
and adding the text data with the text attribute added into a preset image template to obtain the image expansion data.
3. The method of claim 1, wherein determining at least one image to be identified corresponding to the cartridge image comprises:
Inputting the medicine box image into a text detection model for text detection to obtain text position information corresponding to the medicine box image;
And cutting the medicine box image according to the text position information to obtain at least one image to be identified.
4. A drug name recognition method according to claim 3, wherein the text detection model comprises at least a feature extraction layer; inputting the medicine box image into a text detection model for text detection to obtain text position information corresponding to the medicine box image, wherein the text detection comprises the following steps:
inputting the medicine box image into the feature extraction layer for feature extraction to obtain a feature image corresponding to the medicine box image;
Determining a binarization feature map corresponding to the feature image;
And determining a text region in the binarization feature map, and determining the text position information according to the text region.
5. The method for identifying a drug name according to any one of claims 1 to 4, further comprising, after the obtaining of the drug name corresponding to the drug cassette image:
and displaying the medicine name and the medicine instruction information.
6. A medicine name recognition apparatus for performing the medicine name recognition method according to any one of claims 1 to 5, the medicine name recognition apparatus comprising:
The image data acquisition module is used for acquiring image training data, wherein the image training data is obtained by extracting texts from sample medicine box images, and acquiring image expansion data, the image expansion data is text data containing medicine information, and the text data is texts with different text lengths, different text directions, different font sizes and different font styles;
The model training module is used for inputting the image training data and the image expansion data into a medicine name recognition model for iterative training until the medicine name recognition model converges to obtain the trained medicine name recognition model;
The image extraction module is used for acquiring a medicine box image to be subjected to medicine name recognition and determining at least one image to be recognized corresponding to the medicine box image;
The medicine name recognition module is used for inputting each image to be recognized into the trained medicine name recognition model to carry out medicine name recognition, and obtaining a medicine name recognition result corresponding to the medicine box image;
And the medicine name fuzzy matching module is used for carrying out medicine name fuzzy matching on the medicine name recognition result based on a preset medicine name information base and an edit distance algorithm to obtain a medicine name corresponding to the medicine box image.
7. A computer device, the computer device comprising a memory and a processor;
The memory is used for storing a computer program;
the processor for executing the computer program and for implementing the drug name recognition method according to any one of claims 1 to 5 when the computer program is executed.
8. A computer readable storage medium, characterized in that the computer readable storage medium stores a computer program which, when executed by a processor, causes the processor to implement the drug name identification method according to any one of claims 1 to 5.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110486506.8A CN113205047B (en) | 2021-04-30 | 2021-04-30 | Medicine name identification method, device, computer equipment and storage medium |
| PCT/CN2021/097413 WO2022227218A1 (en) | 2021-04-30 | 2021-05-31 | Drug name recognition method and apparatus, and computer device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110486506.8A CN113205047B (en) | 2021-04-30 | 2021-04-30 | Medicine name identification method, device, computer equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113205047A CN113205047A (en) | 2021-08-03 |
| CN113205047B true CN113205047B (en) | 2024-05-10 |
Family
ID=77028529
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110486506.8A Active CN113205047B (en) | 2021-04-30 | 2021-04-30 | Medicine name identification method, device, computer equipment and storage medium |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN113205047B (en) |
| WO (1) | WO2022227218A1 (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113642563A (en) * | 2021-08-31 | 2021-11-12 | 平安医疗健康管理股份有限公司 | Drug use rechecking method, device, equipment and storage medium |
| CN114062393A (en) * | 2021-11-17 | 2022-02-18 | 宜宾显微智能科技有限公司 | Microscopic character recognition and defect detection system and method based on deep learning |
| CN114764444A (en) * | 2022-04-06 | 2022-07-19 | 云从科技集团股份有限公司 | Image generation and sample image expansion method, device and computer storage medium |
| CN116168410B (en) * | 2023-04-21 | 2023-06-30 | 江苏羲辕健康科技有限公司 | Medicine box information identification method and system based on neural network |
| CN116740688B (en) * | 2023-08-11 | 2023-11-07 | 武汉市中西医结合医院（武汉市第一医院） | Medicine identification method and system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107818289A (en) * | 2016-09-13 | 2018-03-20 | 北京搜狗科技发展有限公司 | A kind of prescription recognition methods and device, a kind of device for prescription identification |
| CN109471944A (en) * | 2018-11-12 | 2019-03-15 | 中山大学 | Training method, device and the readable storage medium storing program for executing of textual classification model |
| CN109993165A (en) * | 2019-03-28 | 2019-07-09 | 永康市几米电子科技有限公司 | The identification of tablet plate medicine name and tablet plate information acquisition method, device and system |
| WO2019201141A1 (en) * | 2018-04-18 | 2019-10-24 | Beijing Didi Infinity Technology And Development Co., Ltd. | Methods and systems for image processing |
| CN111126140A (en) * | 2019-11-19 | 2020-05-08 | 腾讯科技（深圳）有限公司 | Text recognition method and device, electronic equipment and storage medium |
| WO2020221298A1 (en) * | 2019-04-30 | 2020-11-05 | 北京金山云网络技术有限公司 | Text detection model training method and apparatus, text region determination method and apparatus, and text content determination method and apparatus |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111488826B (en) * | 2020-04-10 | 2023-10-17 | 腾讯科技（深圳）有限公司 | Text recognition method and device, electronic equipment and storage medium |
-
2021
- 2021-04-30 CN CN202110486506.8A patent/CN113205047B/en active Active
- 2021-05-31 WO PCT/CN2021/097413 patent/WO2022227218A1/en active Application Filing
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107818289A (en) * | 2016-09-13 | 2018-03-20 | 北京搜狗科技发展有限公司 | A kind of prescription recognition methods and device, a kind of device for prescription identification |
| WO2019201141A1 (en) * | 2018-04-18 | 2019-10-24 | Beijing Didi Infinity Technology And Development Co., Ltd. | Methods and systems for image processing |
| CN109471944A (en) * | 2018-11-12 | 2019-03-15 | 中山大学 | Training method, device and the readable storage medium storing program for executing of textual classification model |
| CN109993165A (en) * | 2019-03-28 | 2019-07-09 | 永康市几米电子科技有限公司 | The identification of tablet plate medicine name and tablet plate information acquisition method, device and system |
| WO2020221298A1 (en) * | 2019-04-30 | 2020-11-05 | 北京金山云网络技术有限公司 | Text detection model training method and apparatus, text region determination method and apparatus, and text content determination method and apparatus |
| CN111126140A (en) * | 2019-11-19 | 2020-05-08 | 腾讯科技（深圳）有限公司 | Text recognition method and device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113205047A (en) | 2021-08-03 |
| WO2022227218A1 (en) | 2022-11-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113205047B (en) | Medicine name identification method, device, computer equipment and storage medium | |
| CN111476284B (en) | Image recognition model training and image recognition method and device and electronic equipment | |
| CN109117777B (en) | Method and device for generating information | |
| US20210295114A1 (en) | Method and apparatus for extracting structured data from image, and device | |
| CN111027563A (en) | Text detection method, device and recognition system | |
| CN111488826A (en) | Text recognition method and device, electronic equipment and storage medium | |
| CN112287914B (en) | PPT video segment extraction method, device, equipment and medium | |
| WO2022089170A1 (en) | Caption area identification method and apparatus, and device and storage medium | |
| US11856277B2 (en) | Method and apparatus for processing video, electronic device, medium and product | |
| CN112396049A (en) | Text error correction method and device, computer equipment and storage medium | |
| CN113901954A (en) | Document layout identification method and device, electronic equipment and storage medium | |
| CN113159013B (en) | Paragraph identification method, device, computer equipment and medium based on machine learning | |
| CN111859093A (en) | Sensitive word processing method and device and readable storage medium | |
| CN114596566A (en) | Text recognition method and related device | |
| WO2021237227A1 (en) | Method and system for multi-language text recognition model with autonomous language classification | |
| CN118015644B (en) | Social media keyword data analysis method and device based on pictures and characters | |
| CN112395834B (en) | Brain graph generation method, device and equipment based on picture input and storage medium | |
| CN112839185B (en) | Method, apparatus, device and medium for processing image | |
| CN111986259B (en) | Training of pigment and text detection model, auditing method of video data and related device | |
| CN113537187A (en) | Text recognition method and device, electronic equipment and readable storage medium | |
| CN115130437B (en) | Intelligent document filling method and device and storage medium | |
| CN116774973A (en) | Data rendering method, device, computer equipment and storage medium | |
| CN115858776A (en) | Variant text classification recognition method, system, storage medium and electronic equipment | |
| WO2022105120A1 (en) | Text detection method and apparatus from image, computer device and storage medium | |
| CN113486171A (en) | Image processing method and device and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |