An intelligent generation and matching method, system and medium for image standard inspection items

Abstract

The present invention discloses an intelligent generation and matching method, system and medium for imaging standard inspection items. The method includes establishing a standard series library of imaging inspection items and generating an imaging standard item library; generating matching corpus and corresponding common functions of the matching corpus and saving them to the corpus library; obtaining the name of the inspection item to be matched input by the user; performing training and algorithm matching of multiple machine learning models on the name of the inspection item to be matched and the data in the corpus library to obtain the best-matched inspection item. The system includes an inspection item standard generation module and an inspection item standard matching module. The present invention combines information such as inspection equipment, equipment type, inspection site, scanning method, post-processing, and positioning contained in the inspection item to generate a standard format for the naming and coding of inspection items, as well as functions of matching and mapping non-standard inspection items to standard names using a variety of artificial intelligence matching and recommendation algorithms.

Description

Technical Field

[0001] The present invention relates to the field of medical technologies, and particularly to a method, system and medium for intelligently generating and matching imaging standard examination items. Background Art

[0002] At present, in the medical imaging industry, there is no unified standard for naming and standardizing examination item names, which brings inconvenience and increased costs to the intercommunication of medical institutions, doctor-patient communication and the statistics of data materials. Summary of the Invention

[0003] In view of the defects in the prior art, the present invention provides a method, system and medium for intelligently generating and matching imaging standard examination items;

[0004] In a first aspect, an imaging standard examination item intelligent generation and matching method is characterized by comprising the following steps:

[0005] S1: Establish a standard series library of imaging examination items, including hierarchical parts, equipment and types, scanning and post-processing, positioning, aliases, etc., and combine them according to industry standard specifications to generate CT, MR nuclear magnetic resonance, DR, NM nuclear medicine, US ultrasound imaging standard item libraries;

[0006] S2: Read the original data from the standard series library of imaging examination items and the imaging standard item library, generate and save a matching corpus, and create a common function for reading the matching corpus;

[0007] S3: Obtain the information of the examination item to be matched input by the user, and perform training and algorithm matching of multiple machine learning models on the information of the examination item to be matched and the data in the corpus to obtain the best-matched examination item;

[0008] S4: Obtain the query condition input by the user, extract the keywords in the query condition, and query the imaging standard item library to obtain the best standard examination item.

[0009] Preferably, the specific method of step S3 is:

[0010] S31: Obtain the name of the examination item to be matched from the information of the examination item to be matched, and obtain the standard item name from the corpus;

[0011] S32: Compare the similarity between the name of the examination item to be matched and the standard item name at the character level to obtain a character-level matching result;

[0012] S32: Compare the similarity between the name of the examination item to be matched and the standard item name at the word level to obtain a word-level matching result;

[0013] S33: Combine the character-level matching and word-level matching results, perform combined score weighting processing, and sort according to the scores to obtain the best-matched examination item.

[0014] In a second aspect, an intelligent generation and matching system for imaging standard examination items includes: an examination item standard generation module, configured to combine information from an imaging examination standard series library such as examination equipment, equipment type, examination site, scanning method, post-processing, and positioning included in the examination items, generate an imaging standard item library, and display the imaging standard item library; an examination item standard matching module, configured to map non-standard examination items to the standard item names and standard codes in the imaging standard item library by using multiple artificial intelligence matching and recommendation algorithms.

[0015] Preferably, the examination item standard generation module includes: an imaging examination standard series library establishment unit, configured to create an imaging examination standard series library, where the standard series library includes a first-level site library, a second-level site library, a third-level site library, a standard equipment library, an equipment type library, a standard scan, a post-processing library, a standard positioning library, an alias library, and other additional spare word libraries; a standard generation unit, configured to extract data from the standard series library and generate an imaging standard item library according to industry standard specifications, where the imaging standard item library includes a CT standard item library, an MR standard item library, a DR standard item library, an NM standard item library, and a US standard item library; an imaging standard item library display unit, configured to display information related to standard examination items in the imaging standard item library, where the information related to the standard examination items includes standard examination item names, standard examination item codes, standard sites, standard equipment, equipment types, scanning methods, post-processing, and positioning data.

[0016] Preferably, the examination item standard matching module includes: a matching corpus generation and saving unit, configured to read original data from the imaging standard item library, generate a matching corpus, and save it as a pickle file; a corpus reading common function unit, configured to establish and save a corpus, and create a common function for reading the matching corpus;

[0017] an algorithm matching unit, configured to read corpus data and perform multiple machine learning model training and algorithm matching with the information of the to-be-matched examination items input by the user to obtain the best-matched examination items; among them, the examination item information includes the name of the to-be-matched examination item and the examination equipment; a conditional combination matching standard name unit, configured to obtain keywords of the information of the to-be-matched examination items input by the user, and query standard examination items based on the keyword conditions.

[0018] Preferably, the algorithm matching unit includes: a character-level matching unit for comparing the similarity between the name of the item to be matched and the standard item at the character granularity to obtain the best-matched item; a word-level matching unit for comparing the similarity between the name of the item to be matched and the standard item name at the word granularity to obtain the best-matched item; and a matching result combination unit for combining the character-level matching and word-level matching results, performing a score weighting process to obtain the final matching result. The score weighting process includes: the score ratio allocation of character-level matching, word-level TF-IDF matching, and word-level Word2vec matching, and special weighting processing for the inspection part.

[0019] Preferably, the character-level matching unit includes an input information processing unit I, an inspection part matching unit, and an inspection item name matching unit; the word-level matching unit includes an input information processing unit II, a TF-IDF matching unit, and a word2vec matching unit.

[0020] Preferably, the inspection item name matching unit includes an input information and matching result information unit and a matching method unit.

[0021] In a third aspect, an electronic device includes a processor, an input device, an output device, and a memory. The processor, input device, output device, and memory are interconnected. The memory is used to store a computer program, and the computer program includes program instructions. The processor is configured to execute any one of the methods in claims 1-3 when calling the program instructions.

[0022] In a fourth aspect, a computer-readable storage medium stores a computer program, and the computer program includes program instructions. The program instructions execute any one of the above methods when running on a computer.

[0023] The beneficial effects of the present invention are as follows: The present invention can realize information such as the inspection devices, device types, inspection parts, scanning methods, post-processing, and positioning included in the combined inspection items, generate a standard format for naming and coding medical imaging inspection items, and include functions of matching and mapping non-standard inspection items to standard names using a variety of artificial intelligence matching and recommendation algorithms. The technical improvements therein include the use of artificial intelligence character-level matching algorithms such as the longest common subsequence similarity calculation, edit distance similarity calculation, rule keyword weighting, and exact matching weighting of part names; word-level matching TF-IDF (term frequency-inverse document frequency), word2vec word vector training neural network model to calculate the sentence vector distance similarity; reasonably adjusting the parameters to combine the character-level matching result set and the word-level matching result set based on the principle of balancing the coverage and accuracy of inspection item matching, and making reasonable score combination weighting for the ratio of the weights of medical imaging inspection parts and inspection methods, etc., which overcomes the problem that the query results of standard inspection items that cannot recognize different orders, different combinations, different weights, and different combinations of keywords with different meanings by simple fuzzy queries; the data involved covers professional terms in the field of medical imaging inspection; and an industry standard for naming medical imaging inspection items is formulated. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the description of the specific embodiments or the prior art. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0025] Figure 1 It is a flowchart of a method for intelligent generation and matching of medical imaging inspection items provided in Embodiment 1 of the present invention;

[0026] Figure 2 It is a system block diagram of a system for intelligent generation and matching of medical imaging inspection items provided in Embodiment 2 of the present invention;

[0027] Figure 3 It is a schematic structural diagram of an electronic device provided in Embodiment 3 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0028] The following will describe in detail the embodiments of the technical solutions of the present invention with reference to the drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention, so they are only examples and cannot be used to limit the protection scope of the present invention.

[0029] It should be noted that unless otherwise specified, the technical terms or scientific terms used in this application should have the ordinary meaning understood by those skilled in the art to which the present invention belongs.

[0030] Embodiment 1

[0031] Reference Figure 1 , an embodiment of the present invention provides an intelligent generation and matching method for imaging standard inspection items, including the following steps:

[0032] S1: Establish a standard series library for imaging inspection items, including grading parts, equipment and types, scanning and post-processing, positioning, aliases, etc., and combine them according to industry standard specifications to generate CT, MR nuclear magnetic resonance, DR, NM nuclear medicine, and US ultrasound imaging standard item libraries;

[0033] S2: Read the original data from the standard series library of imaging inspection items and the imaging standard item library, generate and save the matching corpus, and create a common function for reading the matching corpus;

[0034] S3: Obtain the information of the inspection item to be matched input by the user, and perform training and algorithm matching of multiple machine learning models on the information of the inspection item to be matched and the data in the corpus to obtain the best matching inspection item;

[0035] S4: Obtain the query condition input by the user, extract the keywords in the query condition, and query the imaging standard item library to obtain the best standard inspection item.

[0036] According to the user's input, different results can be obtained. If the information of the inspection item to be matched is input by the user, the best matching inspection item is obtained. If the input is a query condition, the best standard inspection item is obtained. The algorithm matching method and the traditional library table query method are respectively adopted, which can meet the user's all-round needs.

[0037] Among them, the specific method of step S3 is:

[0038] S31: Obtain the name of the inspection item to be matched from the information of the inspection item to be matched, and obtain the standard item name from the corpus;

[0039] S32: Use the character-level matching model to compare the similarity between the name of the inspection item to be matched and the standard item name at the character level to obtain the character-level matching result;

[0040] S32: Use the word-level matching model to compare the similarity between the name of the inspection item to be matched and the standard item name at the word level to obtain the word-level matching result;

[0041] S33: Combine the character-level matching and word-level matching results, perform combined score weighting processing, and sort according to the scores to obtain the best matching inspection item.

[0042] Specifically, the specific method of using the character-level matching model includes the input information processing method, the inspection part matching unit method, and the inspection item name matching method to obtain the best matching inspection item:

[0043] The input information processing method maps the third-level parts to the standard names of the second-level inspection parts, maps the aliases to the standard names, formulates rules for classifying and cross-referencing problems, composite item identifiers, punctuation, English letter cases, and specific part inspection conventions for the third-level parts not covered by the standard library and alias library, combines and filters the input information, splits the input information filtered by the rules, and extracts the information required for the standard library fields from the standard library: equipment, equipment type, scanning method, post-processing, and positioning.

[0044] The inspection part matching method extracts the equipment, equipment type, scanning method, post-processing, and positioning keywords from the input information, and then uses the remaining information to match with the CT / MR / DR second-level standard inspection part corpus. Methods such as the longest common subsequence similarity calculation, edit distance similarity calculation, rule keyword weighting, and exact part name matching weighting are used for matching, and the best matching inspection part result is obtained by sorting according to the matching degree score.

[0045] The inspection item name matching method sends the inspection part, inspection method, positioning, and other information into the matching algorithm. By using the longest common subsequence similarity calculation, edit distance similarity calculation, rule keyword weighting, and exact part name matching weighting, the character-level similarity scores with each entry of the standard inspection items are calculated and sorted to obtain the rough screening recommended result set of similarity matching. Among them, the longest common subsequence similarity calculation is to extract the cumulative value feature of the length of the longest identical string contained in both matches, and measure the similarity with the length of the repeated words in the sentence. For example: A: "right humerus" vs "left humerus" counts 2; B: "right humerus", "left femur" counts 1. The similarity is measured by the maximum repeated word count, and group A with more repeated words is more similar than group B. Among them, the edit distance similarity calculation is to extract the transformation complexity of both matches, and measure the similarity features of the sentence length, sentence order, and repeated word dimension with the transformation complexity of both sides. For example: A: "right knee" vs "left knee"; B: "right knee", "left knee patella". Compared with group A, group B needs to transform 3 words to be consistent, while group A only needs to transform 1 word, which is more complex. In the case of the same longest common subsequence count, group B has a higher complexity, so group A is determined to be more similar based on the lower complexity. Among them, the rule keyword weighting means that in the similar features, the words related to the part are weighted. For example: using the longest common subsequence count, A: "right humerus" vs "left humerus" counts 2; B: "right humerus" vs "right femur" counts 2. In the case of the same count, group A has the same humerus, while group B has different parts. Then the core part word is weighted to reduce the weight of the weak related word "right". Among them, the exact part name matching weighting means that in the similar features, the exact part name is set with the highest weight. For example: "right humerus" vs "left humerus", in the case of directly matching the part name "right humerus", "right humerus" is set with the highest weight for exact matching.

[0046] The specific methods of the word-level matching model include the input information processing method, the TF-IDF matching model, and the Word2Vec matching model:

[0047] The input information processing method maps the third-level parts to the standard names of the second-level examination parts, maps the aliases to the standard names, formulates rules for the classification and cross-association problems, composite item identifiers, punctuation, English letter cases, and specific part examination conventions that are not covered by the standard library and alias library, and combines filtering and standardizes the input information;

[0048] The TF-IDF matching model reads the CT / MR / DR standard name matching corpus, performs word segmentation, creates a word list vector, inputs the sentence to be matched for word segmentation and creates a word vector representation, calculates the similarity between the input sentence vector and the traversed standard name vectors, and sorts to obtain the rough screening TF-IDF inspection item matching result set; among them, the TF-IDF term frequency-inverse document frequency model principle is to generate a term frequency word vector based on statistical thinking for vector similarity calculation;

[0049] The Word2Vec matching model reads the CT / MR / DR standard name matching corpus, performs word segmentation, trains a machine learning model to generate word vectors, calculates sentence vectors, inputs the sentence to be matched for word segmentation and creates a word vector representation, calculates the Euclidean distance and cosine distance between the input sentence and the traversed standard name sentence vectors, and sorts to obtain the rough screening word2vec inspection item matching result set; among them, the Word2Vec model is trained to generate word vectors based on the neural network machine learning idea, used to calculate the combined sentence vector, and calculates the sentence vector distance to obtain the similarity score of the two sentences;

[0050] The method of combining multiple matching result weight processing refers to reasonably adjusting the parameters to combine the character-level matching result set and the word set matching result set in principle of balancing the two directions of inspection item matching coverage and accuracy, and making a reasonable score combination weighting for the ratio of the medical imaging examination part weight to the examination method weight, so that the imaging part examination weight is greater than the examination method and positioning information weights, with the aim of improving the practical significance of the final inspection item name matching result with the part as the focus.

[0051] The above is the description of an intelligent generation and matching method for imaging standard examination items provided by the embodiments of the present invention.

[0052] An embodiment of the present invention provides a method for generating and matching medical imaging examination items, which can combine information such as examination equipment, equipment type, examination site, scanning method, post-processing, and positioning included in the examination items to generate a standard format for naming and coding examination items, as well as include functions of matching and mapping non-standard examination items to standard names using a variety of artificial intelligence matching and recommendation algorithms. The technical improvements therein include the use of artificial intelligence character-level matching algorithms such as the longest common subsequence similarity calculation, edit distance similarity calculation, rule keyword weighting, and accurate matching weighting of part names; word-level matching TF-IDF (term frequency-inverse document frequency), word2vec word vector training neural network model to calculate the sentence vector distance similarity; reasonably adjusting the parameters to combine the character-level matching result set and the word set matching result set based on the principle of balancing the coverage and accuracy of examination item matching, and making reasonable score combination weighting for the ratio of the weight of medical imaging examination sites and the weight of examination methods. It overcomes the problem that the query results of standard examination items with different orders, different combinations, different weights, and different combinations of keywords with different meanings cannot be recognized by simple fuzzy queries. The data involved covers professional terms in the field of medical imaging examinations; it formulates industry specifications for the names of imaging examination items.

[0053] Embodiment 2

[0054] As Figure 2 As described above, an embodiment of the present invention provides an intelligent generation and matching system for imaging standard examination items, including: an examination item standard generation module, which is used to combine information in the imaging examination standard series library such as examination equipment, equipment type, examination site, scanning method, post-processing, and positioning included in the examination items to generate an imaging standard item library and display the imaging standard item library; an examination item standard matching module, which is used to map non-standard examination items to standard item names and standard codes in the imaging standard item library using a variety of artificial intelligence matching and recommendation algorithms.

[0055] In this embodiment, the examination item standard generation module includes a unit for establishing an imaging examination standard series library, a unit for generating imaging standard items, and a unit for displaying the imaging standard item library; the unit for establishing an imaging examination standard series library is used to create an imaging examination standard series library, including a first-level part library, a second-level part library, a third-level part library, a standard equipment library, an equipment type library, a standard scan, a post-processing library, a standard positioning library, an alias library, and other additional spare word libraries; the standard generation unit is used to extract data from the standard series library and generate CT, MR nuclear magnetic resonance, DR, NM nuclear medicine, and US ultrasound imaging standard item libraries according to industry standard specifications; the standard library display unit is used to display the generated examination item standard library, standard parts, standard equipment, standard examinations, and positioning.

[0056] It should be noted that the inspection item standard matching module includes a unit for generating and saving matching corpus, a common function unit for reading the corpus, an algorithm matching unit, and a conditional combination matching standard name unit. The unit for generating and saving matching corpus is used to read the original data from the standard library, generate the format required by the matching algorithm, and save it as a pickle file. The common function unit for reading the corpus is used to establish a common function for reading the corpus to read the saved matching corpus. The algorithm matching unit is used to read the corpus data and the input information for training and algorithm matching with multiple machine learning models to obtain the best matching inspection items. The conditional combination matching standard name unit is used to query the standard inspection items by keyword conditions.

[0057] It should be noted that the algorithm matching unit includes a character-level matching unit, a word-level matching unit, and a matching result combination unit. The character-level matching unit is used to compare the similarity between the input inspection item name and the standard item name at the character granularity to obtain the best matching inspection item. The word-level matching unit is used to compare the similarity between the input inspection item name and the standard item name at the word granularity to obtain the best matching inspection item. The matching result combination unit is used to combine the character-level matching and word-level matching results, perform the addition of part scores only, and obtain the final matching result.

[0058] It should be noted that the character-level matching unit includes an input information processing unit 1, an inspection part matching unit, and an inspection item name matching unit. The input information unit 1 is used to process the mapping of the third-level part to the standard name of the second-level inspection part, map the alias to the standard name, formulate rules for the classification and cross-association problems of the third-level parts not covered by the standard library and alias library, compound item identification, punctuation, case of English letters, specific part inspection conventions, etc., perform combined filtering and standardize the input information, split the input information filtered by the rules, and extract the information required by the standard library fields from the standard library: equipment, equipment type, scanning method, post-processing, and positioning. The inspection part matching unit is used to match the remaining information after extracting the keywords of equipment, equipment type, scanning method, post-processing, and positioning from the input information with the CT / MR / DR second-level standard inspection part corpus, and sort and screen according to the matching degree score to obtain the best matching inspection part result. The inspection item name matching unit is used to match the matching inspection part result and the scanning method combination input to the character-level matching model to obtain the character-level standard inspection item matching result set. The inspection item name matching unit includes an input information and matching result information unit and a matching method unit. The input information and matching result information unit is used to create a combination of the inspection part matching result, the filtered and standardized scanning method, post-processing, positioning, and other key information that needs to be included in the standard name as the input small set, and screen the matching result small set: roughly screen the small set result set of CT / MR / DR standard item names containing the input part and scanning, post-processing, and positioning information, and process the result set rules of the non-part fixed items.

[0059] It should be noted that the matching method unit is used to send information such as the examination site, examination method, and positioning to the matching algorithm. By using the longest common subsequence similarity calculation, edit distance similarity calculation, rule keyword weighting, and exact matching weighting of part names, the character-level similarity scores of each item of the standard examination items are calculated, sorted, and a rough screening recommendation result set of similarity matching is obtained; the word-level matching unit includes an input information processing unit II, a TF-IDF matching unit, and a word2vec matching unit; the input information unit II is used to process the standard names of mapping the third-level parts to the second-level examination parts, mapping the aliases to the standard names, formulating rules for the classification and cross-association problems, compound item identification, punctuation, case of English letters, and specific part examination conventions that are not covered by the standard library and alias library, and filtering and standardizing the input information through combination; the TF-IDF matching unit is used to read the CT / MR / DR standard name matching corpus, perform word segmentation, create a word list vector, perform word segmentation on the sentence to be matched, create a word vector representation, calculate the similarity between the input sentence vector and the traversed standard name vector, and sort to obtain a rough screening TF-IDF examination item matching result set; the word2vec matching unit is used to read the CT / MR / DR standard name matching corpus, perform word segmentation, train a machine learning model to generate word vectors, calculate sentence vectors, perform word segmentation on the sentence to be matched, create a word vector representation, calculate the Euclidean distance and cosine distance between the input sentence and the traversed standard name sentence vectors, and sort to obtain a rough screening word2vec examination item matching result set.

[0060] The intelligent generation and matching system for standard examination items of artificial intelligence medical images provided by the present invention and the intelligent generation and matching system for standard examination items of artificial intelligence medical images provided in Embodiment 1 are based on the same inventive concept and have the same beneficial effects, which will not be elaborated here.

[0061] Embodiment 3

[0062] As Figure 3 shown, the embodiment of the present invention also provides a schematic structural diagram of an electronic device. The device includes a processor, an input device, an output device, and a memory. The processor, input device, output device, and memory are interconnected. The memory is used to store a computer program, and the computer program includes program instructions. The processor is configured to call the program instructions to execute the method described in the second embodiment above.

[0063] It should be understood that in the embodiments of the present invention, the so-called processor may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0064] The input device may include a touchpad, a fingerprint sensor (for collecting the fingerprint information and the direction information of the fingerprint of the user), a microphone, etc., and the output device may include a display (such as an LCD), a speaker, etc.

[0065] The memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A part of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0066] In a specific implementation, the processor, input device, and output device described in the embodiments of the present invention may implement the implementation manners described in the method embodiments provided by the embodiments of the present invention, or may also implement the implementation manners of the system embodiments described in the embodiments of the present invention, which will not be elaborated herein.

[0067] Embodiment 3

[0068] The embodiments of the present invention provide an embodiment of a computer-readable storage medium. The computer storage medium stores a computer program, and the computer program includes program instructions. When the program instructions are executed by a processor, the processor is caused to execute the method described in the above embodiments.

[0069] The computer-readable storage medium may be an internal storage unit of the terminal described in the foregoing embodiments, such as the hard disk or memory of the terminal. The computer-readable storage medium may also be an external storage device of the terminal, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc. equipped on the terminal. Further, the computer-readable storage medium may also include both the internal storage unit and the external storage device of the terminal. The computer-readable storage medium is used to store the computer program and other programs and data required by the terminal. The computer-readable storage medium may also be used to temporarily store the data that has been output or is to be output.

[0070] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been generally described according to their functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0071] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described terminal and unit can refer to the corresponding processes in the foregoing method embodiments, and will not be described herein again.

[0072] In several embodiments provided in the present application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed couplings or direct couplings or communication connections to each other may be indirect couplings or communication connections through some interfaces, devices or units, and may also be electrical, mechanical or other forms of connection.

[0073] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the various embodiments of the present invention, and they should all be covered by the scope of the claims and the specification of the present invention.

Claims

1. An intelligent generation and matching method for imaging standard examination items, characterized in that, it includes the following steps: S1: Establish a standard series library of imaging examination items, including hierarchical parts, equipment and types, scanning and post-processing, positioning, aliases, and combine them according to industry standard specifications to generate CT, MR nuclear magnetic resonance, DR, NM nuclear medicine, and US ultrasound imaging standard item libraries; S2: Read the original data from the standard series library of imaging examination items and the imaging standard item library, generate and save the matching corpus, and create a common function for reading the matching corpus; S3: Obtain the name of the examination item to be matched input by the user; perform training and algorithm matching of multiple machine learning models on the name of the examination item to be matched and the data in the corpus to obtain the best-matched examination item; S4: Obtain the query conditions input by the user, extract the keywords in the query conditions to query the imaging standard item library, and obtain the best standard examination item; The specific method for obtaining the best-matched examination item in step S3 is: S31: Obtain the name of the examination item to be matched, and obtain the standard item name from the corpus; S32: Use the character-level matching model to compare the similarity between the name of the examination item to be matched and the standard item name at the character level to obtain the character-level matching result; S32: Use the word-level matching model to compare the similarity between the name of the examination item to be matched and the standard item name at the word level to obtain the word-level matching result; S33: Combine the character-level matching and word-level matching results, perform combined score weighting processing, and sort according to the scores to obtain the best-matched examination item.

2. An intelligent generation and matching system for imaging standard examination items, characterized in that, it includes: An examination item standard generation module, which is used to combine the information of the standard series library of imaging examinations including the examination equipment, equipment type, examination part, scanning method, post-processing, and positioning contained in the examination item, generate an imaging standard item library, and display the imaging standard item library; An examination item standard matching module, which is used to map the non-standard examination items to the standard item names and standard codes in the imaging standard item library by using various artificial intelligence matching and recommendation algorithms; The examination item standard generation module includes: A unit for establishing an imaging examination standard series library, which is used to create an imaging examination standard series library, and the standard series library includes a first-level part library, a second-level part library, a third-level part library, a standard equipment library, an equipment type library, a standard scanning and post-processing library, a standard positioning library, an alias library, and other additional spare word libraries; An imaging standard item library generation unit, which is used to extract the data of the standard series library and generate an imaging standard item library according to industry standard specifications. The imaging standard item library includes a CT standard item library, an MR standard item library, a DR standard item library, an NM nuclear medicine standard item library, and a US ultrasound standard item library; An imaging standard item library display unit, which is used to display the relevant information of the standard examination items in the imaging standard item library. The relevant information of the standard examination items includes the standard examination item name, standard examination item code, standard part, standard equipment, equipment type, scanning method, post-processing, and positioning data; The examination item standard matching module includes: Generate and save matching corpus units for reading raw data from the standard series library of imaging examination items and the imaging standard item library, generating matching corpus and saving it as a pickle file; Corpus reading common function unit for establishing and saving a corpus and creating a common function for reading matching corpus; Algorithm matching unit for reading corpus data and information of the inspection items to be matched input by the user for training multiple machine learning models and algorithm matching to obtain the best-matched inspection items; wherein, the inspection item information includes the name of the inspection item to be matched and the inspection equipment; Condition combination matching standard name unit for obtaining keywords of the inspection item information to be matched input by the user and querying standard inspection items based on the keyword conditions; The algorithm matching unit includes: Character-level matching unit for comparing the similarity between the name of the inspection item to be matched and the standard item name at the character granularity to obtain the best-matched inspection item; Word-level matching unit for comparing the similarity between the name of the inspection item to be matched and the standard item name at the word granularity to obtain the best-matched inspection item; Matching result combination unit for combining the character-level matching and word-level matching results, performing score weighting processing to obtain the final matching result, and the score weighting processing includes: score ratio allocation of character-level matching, word-level TF-IDF matching, word-level Word2vec matching and special weighting processing for the inspection site.

3. An intelligent generation and matching system for imaging standard inspection items according to claim 2, characterized in that, the character-level matching unit includes an input information processing unit I, an inspection site matching unit, and an inspection item name matching unit; the word-level matching unit includes an input information processing unit II, a TF-IDF matching unit, and a word2vec matching unit.

4. An intelligent generation and matching system for imaging standard inspection items according to claim 3, characterized in that, the inspection item name matching unit includes an input information and matching result information unit and a matching method unit.

5. An electronic device, characterized in that, it includes a processor, an input device, an output device and a memory, the processor, input device, output device and memory are interconnected, the memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to execute the method in claim 1 when calling the program instructions.

6. A computer-readable storage medium, characterized in that, the computer-readable storage medium stores a computer program, the computer program includes program instructions, and the program instructions execute the method in claim 1 when running on a computer.

Images