CN109920509B - Drug information identification method, device, computer equipment and storage medium - Google Patents
Drug information identification method, device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN109920509B CN109920509B CN201910042928.9A CN201910042928A CN109920509B CN 109920509 B CN109920509 B CN 109920509B CN 201910042928 A CN201910042928 A CN 201910042928A CN 109920509 B CN109920509 B CN 109920509B
- Authority
- CN
- China
- Prior art keywords
- information
- medicine
- voice
- broadcasting
- usage
- 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 252
- 238000000034 method Methods 0.000 title claims abstract description 37
- 229940079593 drug Drugs 0.000 title claims description 74
- 230000033764 rhythmic process Effects 0.000 claims abstract description 63
- 239000000470 constituent Substances 0.000 claims description 8
- 238000010606 normalization Methods 0.000 claims description 7
- 238000001914 filtration Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 5
- 239000002775 capsule Substances 0.000 description 10
- 238000010586 diagram Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 5
- 230000008901 benefit Effects 0.000 description 4
- 238000012937 correction Methods 0.000 description 4
- 241001672694 Citrus reticulata Species 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 201000010099 disease Diseases 0.000 description 2
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 2
- 230000005802 health problem Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- RZVAJINKPMORJF-UHFFFAOYSA-N Acetaminophen Chemical compound CC(=O)NC1=CC=C(O)C=C1 RZVAJINKPMORJF-UHFFFAOYSA-N 0.000 description 1
- DBAKFASWICGISY-BTJKTKAUSA-N Chlorpheniramine maleate Chemical compound OC(=O)\C=C/C(O)=O.C=1C=CC=NC=1C(CCN(C)C)C1=CC=C(Cl)C=C1 DBAKFASWICGISY-BTJKTKAUSA-N 0.000 description 1
- 239000011358 absorbing material Substances 0.000 description 1
- 238000010521 absorption reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000029142 excretion Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 239000003292 glue Substances 0.000 description 1
- 230000002650 habitual effect Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000002503 metabolic effect Effects 0.000 description 1
- 230000004060 metabolic process Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 201000009240 nasopharyngitis Diseases 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 238000012856 packing Methods 0.000 description 1
- 229960005489 paracetamol Drugs 0.000 description 1
- 230000035790 physiological processes and functions Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Landscapes
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The application relates to the technical field of image recognition, in particular to a medicine information recognition method, a medicine information recognition device, computer equipment and a storage medium, which comprise the following steps: acquiring a medicine information image to be identified, and extracting text information and digital information from the medicine information image to be identified; extracting the medicine name characters in the character information, generating a medicine name after combination, extracting the usage character information in the character information, and combining the usage character information and the digital information to obtain medicine usage information; traversing characters in a preset voice library, acquiring voices to be broadcasted corresponding to the medicine names and the medicine usage information, and broadcasting the voices to be broadcasted according to rhythm parameters corresponding to the medicine names and the medicine usage information. The application effectively realizes automatic identification of medicine information and broadcasting according to language habits of different crowds.
Description
Technical Field
The present application relates to the field of image recognition technologies, and in particular, to a method, an apparatus, a computer device, and a storage medium for identifying drug information.
Background
The medicine is used for preventing, treating and diagnosing diseases. In theory, the medicine refers to chemical substances which can influence the physiological functions of organs and the metabolic activities of cells of the organism, and belongs to the category of medicines. The processes in the body of a drug vary from person to person and many factors can affect absorption, distribution, metabolism, excretion of the drug, thereby affecting the final efficacy. Other factors may also affect the effects of drugs, such as genetics, interactions between drugs, diseases, etc.
Currently, the acquisition of the name and dosage information of the medicine for the patient needs to pass through the doctor's advice or the instruction on the medicine packing box. Often, medicines have a plurality of names, such as trade names, medical names and common names, which cause great trouble to patients when the medicines are used, meanwhile, due to the fact that characters of medical orders are difficult to identify, the usage and metering characters on medicine packages are usually very small, and especially for the old people, when determining how to use the medicines, the old people can only take the medicines according to the language entrust of others. Thus, the potential safety hazard is great, and the damage to the body caused by the mistaking of the medicine by a patient is easy to occur.
Disclosure of Invention
Based on this, it is necessary to provide a drug information identification method, apparatus, computer device, and storage medium for the problem that the name and usage amount cannot be provided to the patient at any time due to the complexity of drug information.
A drug information identification method, comprising the steps of:
acquiring a medicine information image to be identified, and extracting text information and digital information from the medicine information image to be identified;
Extracting the medicine name characters in the character information, generating a medicine name after combination, extracting the usage character information in the character information, and combining the usage character information and the digital information to obtain medicine usage information;
Traversing characters in a preset voice library, acquiring voices to be broadcasted corresponding to the medicine names and the medicine usage information, and broadcasting the voices to be broadcasted according to rhythm parameters corresponding to the medicine names and the medicine usage information.
In one possible embodiment, the acquiring the to-be-identified medicine information image, extracting the text information and the digital information from the to-be-identified medicine information image includes:
When the distance from any image to be identified containing medicine information to an image acquisition device screen is smaller than a preset distance threshold value, acquiring the image to be identified to obtain an original image to be identified;
Carrying out graying treatment on the original image to be identified to obtain a color normalized image to be identified;
scanning the color normalization image to be recognized line by line according to a preset character height, and then obtaining all characters in the color normalization image to be recognized;
and matching the scanned and recognized characters with character information prestored in a database to obtain character information or digital information corresponding to the scanned and recognized characters.
In one possible embodiment, the extracting the medicine name text in the text information, generating the medicine name after combining, extracting the usage text information in the text information, and combining the usage text information and the digital information to obtain the medicine usage information includes:
matching the recognition result of each word in the word information with the words in the word library to obtain the recognition result of the constituent words;
Extracting the drug name feature words in the recognition results of the constituent words, and obtaining the drug names after arranging and combining the drug name feature words according to a preset drug name naming rule;
extracting the usage text information in the recognition result of the constituent words, and combining the usage text information with the digital information to obtain a plurality of medicine usage information sentences;
And calculating the semantic confidence coefficient of each medicine usage information statement, and determining the medicine usage information according to the semantic confidence coefficient.
In one possible embodiment, traversing characters in a preset voice library, obtaining a voice to be broadcasted corresponding to the medicine name and the medicine usage information, broadcasting the voice to be broadcasted according to prosodic parameters corresponding to the medicine name and the medicine usage information, and including:
Traversing characters in a preset voice library to obtain a plurality of original voices corresponding to the medicine names and the medicine usage information;
Acquiring user language use information, and determining that a certain original voice is the voice to be broadcasted according to the user language use information;
acquiring the medicine name and the historical data of word broadcasting in the medicine usage information, acquiring original rhythm parameters according to the historical data, broadcasting the voice to be broadcasted according to the original rhythm parameters, receiving reflected sound waves of the voice to be broadcasted, correcting the original rhythm parameters according to the reflected sound waves, acquiring final rhythm parameters, and broadcasting the voice to be broadcasted according to the final rhythm parameters.
In one possible embodiment, the matching the character identified by the scanning with the character information pre-stored in the database to obtain the text information or the digital information corresponding to the character identified by the scanning includes:
Acquiring connected domains formed by strokes of the characters, and determining the circumscribed rectangle of each connected domain;
Obtaining pixel values of each point in the external rectangle, and dividing the external rectangle into a plurality of sub-blocks according to the pixel values;
Merging the plurality of sub-blocks into a block to be identified according to the aspect ratio of the preset characters;
Comparing the pixel values of each point in the block to be identified with the stroke pixel values of the characters preset in the database to obtain the stroke confidence;
and acquiring strokes corresponding to the maximum value of the stroke confidence coefficient of each block to be identified, and overlapping the strokes to obtain characters or numbers corresponding to the characters in the circumscribed rectangle.
In one possible embodiment, the traversing the text in the preset voice library, obtaining the voice to be broadcasted corresponding to the medicine name and the medicine usage information, and after broadcasting the voice to be broadcasted according to the prosodic parameters corresponding to the medicine name and the medicine usage information, further includes the step of correcting the voice broadcasting content according to the user feedback information, specifically including:
receiving feedback information sent by a user side, extracting problem information characteristic characters contained in the feedback information, and acquiring problem attributes corresponding to the feedback information according to the problem information characteristic characters;
And adjusting the rhythm parameters according to the problem attributes, if the adjusted rhythm parameters reach the requirements of the user end contained in the feedback information, using the adjusted rhythm parameters as rhythm parameters of the medicine information broadcasting, otherwise, acquiring voice information input by a user, and correcting the voice to be broadcasted according to the voice information input by the user until the requirements of the user end are met.
In one possible embodiment, the obtaining the medicine name and the historical data of word broadcasting in the medicine usage information, obtaining an original prosodic parameter according to the historical data, broadcasting the voice to be broadcasted according to the original prosodic parameter, receiving a reflected sound wave of the voice to be broadcasted, correcting the original prosodic parameter according to the reflected sound wave, obtaining a final prosodic parameter, and broadcasting the voice to be broadcasted according to the final prosodic parameter includes:
Acquiring the medicine name and the historical data of broadcasting of words in the medicine usage information, extracting the pitch, the duration and the volume of each word with the largest occurrence number, performing binarization processing on the pitch, the duration and the volume with the largest occurrence number, and splicing to form the original rhythm parameters;
acquiring a preset statement broadcasting speed, and broadcasting the voice to be broadcasted according to the statement broadcasting speed and the original rhythm parameters;
Receiving an original reflected wave of the voice to be broadcasted, which is reflected back from a preset sound reflecting wall, and filtering the original reflected wave to obtain an actual reflected sound wave, wherein the formula of reflected wave filtering is as follows:
Wherein NMSE represents standard deviation, x (n) represents original reflection wavelength, y (n) represents estimated value, and the actual reflection sound wave is obtained after the original reflection wavelength and the standard deviation are subjected to difference;
And performing difference between the actual reflected sound wave and a preset reflected sound wave threshold value, correcting the original rhythm parameters according to the corresponding relation between the difference value and the pitch, the length and the volume to obtain final rhythm parameters, and broadcasting the voice to be broadcasted according to the final rhythm parameters.
A drug information identification device comprising the following modules:
The information extraction module is used for acquiring a medicine information image to be identified and extracting text information and digital information from the medicine information image to be identified;
The information identification module is used for extracting the medicine name characters in the character information, generating a medicine name after combination, extracting the usage character information in the character information, and combining the usage character information with the digital information to obtain medicine usage information;
The information broadcasting module is used for traversing characters in a preset voice library, acquiring voices to be broadcasted corresponding to the medicine names and the medicine usage information, and broadcasting the voices to be broadcasted according to the rhythm parameters corresponding to the medicine names and the medicine usage information.
A computer device comprising a memory and a processor, the memory having stored therein computer readable instructions which, when executed by the processor, cause the processor to perform the steps of the drug information identification method described above.
A storage medium storing computer readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of the drug information identification method described above.
Compared with the existing mechanism, the application has the following advantages:
(1) By effectively identifying the medicine information and performing voice broadcasting to the user in a proper mode, different users can accurately obtain the medicine information, so that the health problem caused by improper medicine use is avoided;
(2) The acquired medicine information is processed, so that the accuracy of medicine information identification is improved;
(3) By carrying out semantic analysis on the identification result, the medicine name and the usage information are effectively obtained, so that a user can be guided to use the medicine correctly;
(4) And the rhythm parameters are adjusted to obtain a medicine information broadcasting mode which accords with the speaking habit of the user.
Drawings
Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the application.
FIG. 1 is a general flow chart of a method for identifying drug information in one embodiment of the application;
FIG. 2 is a schematic diagram of an information extraction process in a drug information identification method according to an embodiment of the present application;
FIG. 3 is a schematic diagram showing an information identification process in a drug information identification method according to an embodiment of the present application;
fig. 4 is a schematic diagram of an information broadcasting process in a drug information identification method according to an embodiment of the present application;
fig. 5 is a block diagram of a drug information identification device according to an embodiment of the present application.
Detailed Description
The present application will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present application more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.
As used herein, the singular forms "a", "an", "the" and "the" are intended to include the plural forms as well, unless expressly stated otherwise, as understood by those skilled in the art. It will be further understood that the terms "comprises" and/or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
Fig. 1 is an overall flowchart of a drug information identification method according to an embodiment of the present application, as shown in fig. 1, and the drug information identification method includes the following steps:
S1, acquiring a medicine information image to be identified, and extracting text information and digital information from the medicine information image to be identified;
Specifically, when the medicine information image to be identified is obtained, a photographing mode can be adopted to extract a commercially available medicine external package image, and the medicine information image can also be extracted from a medicine information database, such as a hospital database or a pharmaceutical factory database. The extracted text information is mainly text information about the health of the user such as 'medicine name', 'usage', 'notice', etc., and the numerical information is mainly specific numbers after the usage, for example, '1' and '3' in 3 times a day.
S2, extracting the medicine name characters in the character information, generating a medicine name after combination, extracting the usage character information in the character information, and combining the usage character information and the digital information to obtain medicine usage information;
Specifically, when the medicine name text is extracted, since general medicines are classified into medical names and common names, it is necessary to distinguish the medical names from the common names after the medicine name text is extracted. The usage words, 2 words of usage, are usually found on the outer package of the medicine, and the specific usage with numbers is searched before and after the 2 words are searched. For example, the usage: 3 times a day, 1 time 1 tablet, etc.
And S3, traversing characters in a preset voice library, acquiring voices to be broadcasted corresponding to the medicine names and the medicine usage information, and broadcasting the voices to be broadcasted according to the rhythm parameters corresponding to the medicine names and the medicine usage information.
Specifically, the pronunciation of all Chinese characters, including Mandarin and dialects, is stored in the voice library, and foreign language voice can be added in the voice library. Since different speech utterances cause users to understand the content, it is necessary to obtain speech utterances, i.e., prosodic parameters, that are habitual to drug name information and drug usage information. The prosodic parameters are mainly pitch and pause interval between each tone.
According to the embodiment, the medicine information is effectively identified, and voice broadcasting is carried out on the user in a proper mode, so that different users can accurately obtain the medicine information, and the health problem caused by improper use of the medicine is avoided.
Fig. 2 is a schematic diagram of an information extraction process in a drug information identification method according to an embodiment of the present application, where as shown in the drawing, S1, a drug information image to be identified is obtained, and text information and digital information are extracted from the drug information image to be identified, including:
S101, when the distance from any image to be identified containing medicine information to an image acquisition device screen is smaller than a preset distance threshold value, acquiring the image to be identified to obtain an original image to be identified;
Specifically, the method comprises the steps of acquiring the information of the to-be-identified medicine image by adopting a CCD camera for two times, then carrying out pixel point verification on the to-be-identified medicine image acquired for two times, and carrying out the next step if the result of the pixel point verification is within a preset error range, otherwise, carrying out acquisition again.
S102, carrying out graying treatment on the original image to be identified to obtain a color normalized image to be identified;
wherein, the color normalization refers to that the color image is subjected to black-and-white processing, and is changed into a black-and-white image.
S103, scanning the color normalization image to be recognized line by line according to a preset character height, and then obtaining all characters in the color normalization image to be recognized;
The preset character height is determined according to the common character size of the medicine package, and can be determined according to the size of the medicine package.
And S104, matching the scanned and recognized characters with character information pre-stored in a database to obtain character information or digital information corresponding to the scanned and recognized characters.
Specifically, when character information is matched, strokes of the character information can be compared, and only when the scanned character information is consistent with character information prestored in a database, the two strokes are matched, otherwise, the two strokes are not matched.
According to the embodiment, the acquired medicine information is processed, so that the accuracy of medicine information identification is improved.
Fig. 3 is a schematic diagram of an information identification process in a method for identifying medicine information in an embodiment of the present application, as shown in the drawing, S2, extracting medicine name characters in the text information, generating a medicine name after combining, extracting usage text information in the text information, and combining the usage text information and the digital information to obtain medicine usage information, where the method includes:
S201, matching the recognition result of each word in the word information with words in a word library to obtain a recognition result of the formed word;
The recognition result of the word is that two or more words can obtain corresponding words in the word library, for example, the recognition result of some two words in the word information is: "use" and "law" may then correspond to terms in the term library: the usage, and the recognition result of the other two words is "glue" and "element", the corresponding words cannot be obtained from the word library.
S202, extracting the drug name feature words in the recognition results of the constituent words, and obtaining the drug names after arranging and combining the drug name feature words according to a preset drug name naming rule;
Specifically, the drug name feature words may include: medical names and trade names, for example: the paracetamol and chlorphenamine maleate capsules are medical names; quick-acting capsules for cold are trade names, and when the quick-acting capsules are arranged and combined, the quick-acting capsules and the quick-acting capsules cannot be combined because the quick-acting capsules and the quick-acting capsules are required to be combined. Meanwhile, the name of "quick-acting capsule for common cold" cannot be combined according to the naming rule of the medicine name.
S203, extracting the usage text information in the recognition result of the constituent words, and combining the usage text information with the digital information to obtain a plurality of medicine usage information sentences;
Specifically, two words of usage or dosage appear on the medicine package, the two words are searched from the recognition result, then the words and numbers near the 2 words of usage or dosage are extracted according to the scanning sequence during OCR recognition, and then the words and numbers are combined to obtain a plurality of medicine usage information sentences.
S204, calculating semantic confidence coefficient of each medicine usage information statement, and determining the medicine usage information according to the semantic confidence coefficient.
Specifically, the calculation formula used in the confidence calculation is:
F＝Z×(P×(1-P))/E；
f is confidence, E is standard deviation of sample mean, Z is total error, and P is the proportion of the target semantic quantity to the total semantic quantity.
If the confidence coefficient is larger than the confidence coefficient threshold value, the statement is the medicine usage information, otherwise, the statement is not the medicine usage information.
According to the embodiment, the identification result is subjected to semantic analysis, so that the medicine name and the usage information are effectively obtained, and a user can be guided to use the medicine correctly.
Fig. 4 is a schematic diagram of an information broadcasting process in a method for identifying drug information according to an embodiment of the present application, where as shown in the drawing, S3 traverses characters in a preset voice library to obtain a voice to be broadcasted corresponding to the drug name and the drug usage information, and broadcasts the voice to be broadcasted according to prosodic parameters corresponding to the drug name and the drug usage information, where the method includes:
S301, traversing characters in a preset voice library, and acquiring a plurality of original voices corresponding to the medicine names and the medicine usage information;
Specifically, the voice library stores voice information corresponding to different words, each word corresponds to at least one voice information, for example, the voice information corresponding to "eating" the word may be "chi" and "qia". Before inquiring the characters in the preset voice library, the names of the medicines need to be converted, for example, the common names corresponding to the medical names of the capsules are quick-acting cold capsules, in patients, only the common names are always known, but the medical names are not known, in the external medicine package or in the prescription prescribed by doctors, the medical names are always written, and the medical names need to be converted into the common names when voice broadcasting is carried out.
S302, acquiring user language use information, and determining that a certain original voice is the voice to be broadcasted according to the user language use information;
The user language usage information mainly refers to what language and dialect the user uses, such as english and chinese, and whether mandarin or cantonese is used in chinese. The user language usage information may be obtained according to the historical data or according to the voice information input by the user, for example, if the pronunciation of the word "eat" input by the user is "qia", the language used by the user may be determined to be the language near Hunan.
S303, acquiring the medicine name and the historical data of word broadcasting in the medicine usage information, obtaining original rhythm parameters according to the historical data, broadcasting the voice to be broadcasted according to the original rhythm parameters, receiving reflected sound waves of the voice to be broadcasted, obtaining final rhythm parameters after correcting the original rhythm parameters according to the reflected sound waves, and broadcasting the voice to be broadcasted according to the final rhythm parameters.
Specifically, the prosodic parameters mainly include pitch, timbre, pause interval. When the reflected sound wave correction is carried out, the reflected sound wave correction is required to be carried out in a closed room, only one of six surfaces of the room is made of the reflected sound wave material, and the other five surfaces are made of the wave absorbing material, so that only sound waves reflected by the surface of the reflected sound wave material are received, then the actual rhythm parameters are obtained according to the sound wave values, the actual rhythm parameters are compared with the original rhythm parameters, and then the original rhythm parameters are corrected by utilizing an error correction function to obtain the final rhythm parameters.
According to the embodiment, the rhythm parameters are adjusted, so that a medicine information broadcasting mode which accords with speaking habits of users is obtained.
In one embodiment, the step S104 of matching the scanned and recognized character with character information pre-stored in a database to obtain text information or digital information corresponding to the scanned and recognized character includes:
Acquiring connected domains formed by strokes of the characters, and determining the circumscribed rectangle of each connected domain;
Specifically, the connected domain formed by each stroke refers to a region where a "factory" is located, for example, a connected domain formed by two strokes of "a" and the circumscribed rectangle of the connected domain refers to a rectangle with "a" as a horizontal side and a length perpendicular to the horizontal side of "a" as a vertical side.
Obtaining pixel values of each point in the external rectangle, and dividing the external rectangle into a plurality of sub-blocks according to the pixel values;
Specifically, when the outer rectangle is segmented into sub-blocks, the size of a sub-block is preset, the outer rectangle is segmented into equal areas, then the pixel value in each sub-block is counted, the sub-block size with the pixel value smaller than the average value is enlarged, the sub-block size with the pixel value larger than the average value is reduced until the pixel values in the sub-blocks are consistent, and a final sub-block is generated.
Merging the plurality of sub-blocks into a block to be identified according to the aspect ratio of the preset characters;
The aspect ratio of the text refers to the ratio of the width to the height of the text area, for example, the aspect ratio of the "mouth" is 1, the aspect ratio of the "country" is 3 to 4, and the block formed by merging the sub-blocks is ensured to be consistent with the preset aspect ratio.
Comparing the pixel values of each point in the block to be identified with the stroke pixel values of the characters preset in the database to obtain the stroke confidence;
When comparing pixel values, two pixel values can be subjected to binarization processing, and then the pixel values are compared, and when the stroke confidence coefficient is obtained, the following formula is adopted:
F＝Z×(2×S/d)/2；
f is confidence, S is standard deviation of pixel value of each point in the block to be identified and pixel value of any one of the strokes in the database, Z is pixel value of any one of the strokes in the database, and d is error maximum. The maximum error value can be obtained according to historical data statistics, and is generally 5%.
And acquiring strokes corresponding to the maximum value of the stroke confidence coefficient of each block to be identified, and overlapping the strokes to obtain characters or numbers corresponding to the characters in the circumscribed rectangle.
When the strokes are overlapped, if two strokes are intersected, the common part is taken as the stroke to be used.
According to the embodiment, the accuracy of character recognition can be improved by carrying out stroke segmentation recognition on the character information.
In one embodiment, the step S3 of traversing the text in the preset voice library to obtain the voice to be broadcasted corresponding to the drug name and the drug usage information, and after broadcasting the voice to be broadcasted according to the prosodic parameters corresponding to the drug name and the drug usage information, further includes a step of correcting the voice broadcasting content according to the user feedback information, and specifically includes:
receiving feedback information sent by a user side, extracting problem information characteristic characters contained in the feedback information, and acquiring problem attributes corresponding to the feedback information according to the problem information characteristic characters;
the question information characters mainly refer to negative characters or characters such as 'none', or the like or adverbs such as 'small', 'large', and the like. These characters may indicate that the user is unable to obtain drug content information based on the current voice broadcast content.
The question attribute refers to an attribute corresponding to the question information, for example, the question attribute corresponding to "small sound" is "volume", "XX word unclear" corresponds to a different prosodic parameter such as a tone of "XX word".
And adjusting the rhythm parameters according to the problem attributes, if the adjusted rhythm parameters reach the requirements of the user end contained in the feedback information, using the adjusted rhythm parameters as rhythm parameters of the medicine information broadcasting, otherwise, acquiring voice information input by a user, and correcting the voice to be broadcasted according to the voice information input by the user until the requirements of the user end are met.
Specifically, the requirement of the user side can be evaluated through the feedback information of the user, namely, if words such as 'no', 'big' and the like do not appear in the feedback information any more, the broadcasting voice is considered to accord with the requirement of the user. According to different question attributes, a certain parameter of the prosodic parameters can be adjusted in a targeted manner. However, the feedback information of the user may be inaccurate, and then the voice of the user needs to be acquired, and the voice of the medicine information is corrected according to the prosodic parameters of the voice of the user.
According to the embodiment, through feedback of the user on the medicine voice broadcasting content, the broadcasting rhythm parameters are adjusted so as to be more in line with language habits of different users, and the user is convenient to obtain correct medicine information.
In one embodiment, the step S303 of obtaining the medicine name and the historical data of word broadcasting in the medicine usage information, obtaining an original prosodic parameter according to the historical data, broadcasting the voice to be broadcasted according to the original prosodic parameter, receiving a reflected sound wave of the voice to be broadcasted, correcting the original prosodic parameter according to the reflected sound wave, obtaining a final prosodic parameter, and broadcasting the voice to be broadcasted according to the final prosodic parameter includes:
Acquiring the medicine name and the historical data of broadcasting of words in the medicine usage information, extracting the pitch, the duration and the volume of each word with the largest occurrence number, performing binarization processing on the pitch, the duration and the volume with the largest occurrence number, and splicing to form the original rhythm parameters;
Specifically, when the pitch, the duration and the volume are spliced into the original prosodic parameters, the pitch can be placed in front, the duration can also be placed in front, and when the binarization processing is performed, different marks can be performed on the pitch, the duration and the volume so as to search for the parameters in the original prosodic parameters respectively.
Acquiring a preset statement broadcasting speed, and broadcasting the voice to be broadcasted according to the statement broadcasting speed and the original rhythm parameters;
Specifically, the preset statement broadcasting speed is obtained according to statistics of historical data, and is generally set according to the situation of the common population of the medicine, for example, if the main population of the medicine is the old over 60 years old, the broadcasting speed is generally slower.
Receiving an original reflected wave of the voice to be broadcasted, which is reflected back from a preset sound reflecting wall, and filtering the original reflected wave to obtain an actual reflected sound wave, wherein the formula of reflected wave filtering is as follows:
Wherein NMSE represents standard deviation, x (n) represents original reflection wavelength, y (n) represents estimated value, and the actual reflection sound wave is obtained after the original reflection wavelength and the standard deviation are subjected to difference;
And performing difference between the actual reflected sound wave and a preset reflected sound wave threshold value, correcting the original rhythm parameters according to the corresponding relation between the difference value and the pitch, the length and the volume to obtain final rhythm parameters, and broadcasting the voice to be broadcasted according to the final rhythm parameters.
Specifically, the difference is attribute-subdivided, that is, whether the difference is caused by pitch, duration or volume, and then targeted correction is performed.
According to the method, the original broadcasting rhythm parameters are effectively adjusted by utilizing the reflected sound waves, so that the broadcasting of the medicine information can meet the user requirements.
In one embodiment, a drug information identification device is provided, as shown in fig. 5, including the following modules:
An information extraction module 51 configured to acquire a drug information image to be identified, and extract text information and digital information from the drug information image to be identified;
the information identifying module 52 is configured to extract the medicine name text in the text information, generate a medicine name after combination, extract the usage text information in the text information, and combine the usage text information with the digital information to obtain medicine usage information;
the information broadcasting module 53 is configured to traverse the text in the preset voice library, obtain the voice to be broadcasted corresponding to the medicine name and the medicine usage information, and broadcast the voice to be broadcasted according to the prosodic parameters corresponding to the medicine name and the medicine usage information.
In one embodiment, a computer device is provided, where the computer device includes a memory and a processor, where computer readable instructions are stored in the memory, and when the computer readable instructions are executed by the processor, the processor is caused to perform the steps of the drug information identification method in the above embodiments.
In one embodiment, a storage medium storing computer readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of the drug information identification method in the above embodiments is presented. Wherein the storage medium may be a non-volatile storage medium.
Those of ordinary skill in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be implemented by a program to instruct related hardware, the program may be stored in a computer readable storage medium, and the storage medium may include: read Only Memory (ROM), random access Memory (RAM, random Access Memory), magnetic or optical disk, and the like.
The technical features of the above-described embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above-described embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The above-described embodiments represent only some exemplary embodiments of the application, in which the description is more specific and detailed, but should not be construed as limiting the scope of the application. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the application, which are all within the scope of the application. Accordingly, the scope of protection of the present application is to be determined by the appended claims.
Claims (9)
1. A method of identifying medication information, comprising:
acquiring a medicine information image to be identified, and extracting text information and digital information from the medicine information image to be identified;
Extracting the medicine name characters in the character information, generating a medicine name after combination, extracting the usage character information in the character information, and combining the usage character information and the digital information to obtain medicine usage information;
Traversing characters in a preset voice library, acquiring voices to be broadcasted corresponding to the medicine names and the medicine usage information, and broadcasting the voices to be broadcasted according to rhythm parameters corresponding to the medicine names and the medicine usage information;
traversing characters in a preset voice library, acquiring voice to be broadcasted corresponding to the medicine name and the medicine usage information, broadcasting the voice to be broadcasted according to rhythm parameters corresponding to the medicine name and the medicine usage information, and comprising the following steps: traversing characters in a preset voice library to obtain a plurality of original voices corresponding to the medicine names and the medicine usage information; acquiring user language use information, and determining that a certain original voice is the voice to be broadcasted according to the user language use information; acquiring the medicine name and the historical data of word broadcasting in the medicine usage information, acquiring original rhythm parameters according to the historical data, broadcasting the voice to be broadcasted according to the original rhythm parameters, receiving reflected sound waves of the voice to be broadcasted, correcting the original rhythm parameters according to the reflected sound waves, acquiring final rhythm parameters, and broadcasting the voice to be broadcasted according to the final rhythm parameters.
2. The method for identifying drug information according to claim 1, wherein the step of obtaining the drug information image to be identified, extracting text information and digital information from the drug information image to be identified, comprises:
When the distance from any image to be identified containing medicine information to an image acquisition device screen is smaller than a preset distance threshold value, acquiring the image to be identified to obtain an original image to be identified;
Carrying out graying treatment on the original image to be identified to obtain a color normalized image to be identified;
scanning the color normalization image to be recognized line by line according to a preset character height, and then obtaining all characters in the color normalization image to be recognized;
and matching the scanned and recognized characters with character information prestored in a database to obtain character information or digital information corresponding to the scanned and recognized characters.
3. The method for identifying medicine information according to claim 2, wherein the extracting the medicine name text in the text information, combining to generate a medicine name, extracting the usage text information in the text information, combining the usage text information and the digital information to obtain medicine usage information, comprises:
matching the recognition result of each word in the word information with the words in the word library to obtain the recognition result of the constituent words;
Extracting the drug name feature words in the recognition results of the constituent words, and obtaining the drug names after arranging and combining the drug name feature words according to a preset drug name naming rule;
extracting the usage text information in the recognition result of the constituent words, and combining the usage text information with the digital information to obtain a plurality of medicine usage information sentences;
And calculating the semantic confidence coefficient of each medicine usage information statement, and determining the medicine usage information according to the semantic confidence coefficient.
4. The method for identifying medicine information according to claim 2, wherein the step of matching the scanned and identified character with character information pre-stored in a database to obtain text information or digital information corresponding to the scanned and identified character comprises the steps of:
Acquiring connected domains formed by strokes of the characters, and determining the circumscribed rectangle of each connected domain;
Obtaining pixel values of each point in the external rectangle, and dividing the external rectangle into a plurality of sub-blocks according to the pixel values;
Merging the plurality of sub-blocks into a block to be identified according to the aspect ratio of the preset characters;
Comparing the pixel values of each point in the block to be identified with the stroke pixel values of the characters preset in the database to obtain the stroke confidence;
and acquiring strokes corresponding to the maximum value of the stroke confidence coefficient of each block to be identified, and overlapping the strokes to obtain characters or numbers corresponding to the characters in the circumscribed rectangle.
5. The method for identifying drug information according to claim 1, wherein the step of traversing characters in a preset voice library to obtain a voice to be broadcasted corresponding to the drug name and the drug usage information, and after broadcasting the voice to be broadcasted according to prosodic parameters corresponding to the drug name and the drug usage information, further comprises the step of correcting voice broadcasting content according to user feedback information, specifically comprises the steps of:
receiving feedback information sent by a user side, extracting problem information characteristic characters contained in the feedback information, and acquiring problem attributes corresponding to the feedback information according to the problem information characteristic characters;
And adjusting the rhythm parameters according to the problem attributes, if the adjusted rhythm parameters reach the requirements of the user end contained in the feedback information, using the adjusted rhythm parameters as rhythm parameters of the medicine information broadcasting, otherwise, acquiring voice information input by a user, and correcting the voice to be broadcasted according to the voice information input by the user until the requirements of the user end are met.
6. The method for identifying drug information according to claim 1, wherein the obtaining the drug name and the historical data of word broadcasting in the drug usage information, obtaining an original prosodic parameter according to the historical data, broadcasting the voice to be broadcasted according to the original prosodic parameter, receiving a reflected sound wave of the voice to be broadcasted, correcting the original prosodic parameter according to the reflected sound wave, obtaining a final prosodic parameter, and broadcasting the voice to be broadcasted according to the final prosodic parameter comprises:
Acquiring the medicine name and the historical data of broadcasting of words in the medicine usage information, extracting the pitch, the duration and the volume of each word with the largest occurrence number, performing binarization processing on the pitch, the duration and the volume with the largest occurrence number, and splicing to form the original rhythm parameters;
acquiring a preset statement broadcasting speed, and broadcasting the voice to be broadcasted according to the statement broadcasting speed and the original rhythm parameters;
Receiving an original reflected wave of the voice to be broadcasted, which is reflected back from a preset sound reflecting wall, and filtering the original reflected wave to obtain an actual reflected sound wave, wherein the formula of reflected wave filtering is as follows:
Wherein NMSE represents standard deviation, x (n) represents original reflection wavelength, y (n) represents estimated value, and the actual reflection sound wave is obtained after the original reflection wavelength and the standard deviation are subjected to difference;
And performing difference between the actual reflected sound wave and a preset reflected sound wave threshold value, correcting the original rhythm parameters according to the corresponding relation between the difference value and the pitch, the length and the volume to obtain final rhythm parameters, and broadcasting the voice to be broadcasted according to the final rhythm parameters.
7. A medicine information identifying apparatus, comprising:
The information extraction module is used for acquiring a medicine information image to be identified and extracting text information and digital information from the medicine information image to be identified;
The information identification module is used for extracting the medicine name characters in the character information, generating a medicine name after combination, extracting the usage character information in the character information, and combining the usage character information with the digital information to obtain medicine usage information;
the information broadcasting module is used for traversing characters in a preset voice library, acquiring voices to be broadcasted corresponding to the medicine names and the medicine usage information, and broadcasting the voices to be broadcasted according to the rhythm parameters corresponding to the medicine names and the medicine usage information;
The information broadcasting module is specifically used for: traversing characters in a preset voice library to obtain a plurality of original voices corresponding to the medicine names and the medicine usage information; acquiring user language use information, and determining that a certain original voice is the voice to be broadcasted according to the user language use information; acquiring the medicine name and the historical data of word broadcasting in the medicine usage information, acquiring original rhythm parameters according to the historical data, broadcasting the voice to be broadcasted according to the original rhythm parameters, receiving reflected sound waves of the voice to be broadcasted, correcting the original rhythm parameters according to the reflected sound waves, acquiring final rhythm parameters, and broadcasting the voice to be broadcasted according to the final rhythm parameters.
8. A computer device comprising a memory and a processor, the memory having stored therein computer readable instructions which, when executed by the processor, cause the processor to perform the steps of the drug information identification method of any of claims 1 to 6.
9. A storage medium storing computer readable instructions which, when executed by one or more processors, cause the one or more processors to perform the steps of the drug information identification method of any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910042928.9A CN109920509B (en) | 2019-01-17 | 2019-01-17 | Drug information identification method, device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910042928.9A CN109920509B (en) | 2019-01-17 | 2019-01-17 | Drug information identification method, device, computer equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109920509A CN109920509A (en) | 2019-06-21 |
| CN109920509B true CN109920509B (en) | 2024-05-14 |
Family
ID=66960440
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910042928.9A Active CN109920509B (en) | 2019-01-17 | 2019-01-17 | Drug information identification method, device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109920509B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112053761A (en) * | 2020-08-12 | 2020-12-08 | 北京左医健康技术有限公司 | Medication guiding method and device based on image recognition and storage medium |
| CN112967787A (en) * | 2021-01-28 | 2021-06-15 | 壹健康健康产业(深圳)有限公司 | Medicine information input method, device, medium and terminal equipment |
| CN112989974A (en) * | 2021-03-02 | 2021-06-18 | 赵宏福 | Text recognition method and device for automatic word segmentation and spelling and storage medium |
| CN113012783A (en) * | 2021-03-18 | 2021-06-22 | 深圳市瑞意博科技股份有限公司 | Medicine rechecking method and device, computer equipment and storage medium |
| CN116168376B (en) * | 2023-04-18 | 2023-07-18 | 苏州大学 | Voice broadcast medicine box recognition device and medicine box recognition method |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106652995A (en) * | 2016-12-31 | 2017-05-10 | 深圳市优必选科技有限公司 | Voice broadcasting method and system for text |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120330665A1 (en) * | 2011-06-03 | 2012-12-27 | Labels That Talk, Ltd | Prescription label reader |
| US11011259B2 (en) * | 2015-09-04 | 2021-05-18 | Walgreen Co. | Automated pharmacy translation engine for prescription medication instructions |
-
2019
- 2019-01-17 CN CN201910042928.9A patent/CN109920509B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106652995A (en) * | 2016-12-31 | 2017-05-10 | 深圳市优必选科技有限公司 | Voice broadcasting method and system for text |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109920509A (en) | 2019-06-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109920509B (en) | Drug information identification method, device, computer equipment and storage medium | |
| Drossos et al. | Clotho: An audio captioning dataset | |
| ES2394726T3 (en) | Automatic extraction of semantic content and generation of a structured document from speech | |
| CA2949782C (en) | Enhancing reading accuracy, efficiency and retention | |
| WO2021174728A1 (en) | Triage data processing method and apparatus, computer device, and storage medium | |
| CN106874643A (en) | Build the method and system that knowledge base realizes assisting in diagnosis and treatment automatically based on term vector | |
| CN106844351B (en) | Medical institution organization entity identification method and device oriented to multiple data sources | |
| CN108491389B (en) | Method and device for training click bait title corpus recognition model | |
| CN110502750A (en) | Disambiguation method, system, equipment and medium during Chinese medicine text participle | |
| AU2020100604A4 (en) | Expert report editor | |
| Parker | Sounding out sonority | |
| CN109509088A (en) | Loan checking method, device, equipment and medium based on micro- Expression Recognition | |
| CN111199801B (en) | Construction method and application of model for identifying disease types of medical records | |
| Bürki et al. | When words collide: Bayesian meta-analyses of distractor and target properties in the picture–word interference paradigm | |
| CN113642563A (en) | Drug use rechecking method, device, equipment and storage medium | |
| Hu et al. | Ovarian toxicity assessment in histopathological images using deep learning | |
| Chung et al. | Formant trajectory patterns of American English/l/produced by adults and children | |
| Skopeteas | The empirical investigation of information structure | |
| Mack | Sentence processing by non-native speakers of English: Evidence from the perception of natural and computer-generated anomalous L2 sentences | |
| Mousikou et al. | Coarticulation across morpheme boundaries: An ultrasound study of past-tense inflection in Scottish English | |
| CN115457586A (en) | Case information extraction method, device, equipment and storage medium | |
| Politzer-Ahles et al. | N400 evidence that the early stages of lexical access ignore knowledge about phonological alternations | |
| Pires et al. | Brand names of Portuguese medication: understanding the importance of their linguistic structure and regulatory issues | |
| CN113705560A (en) | Data extraction method, device and equipment based on image recognition and storage medium | |
| CN110399610B (en) | Processing system of medicine specification |
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 |