CN113851219A - Intelligent diagnosis guiding method based on multi-mode knowledge graph - Google Patents
Intelligent diagnosis guiding method based on multi-mode knowledge graph Download PDFInfo
- Publication number
- CN113851219A CN113851219A CN202111427220.9A CN202111427220A CN113851219A CN 113851219 A CN113851219 A CN 113851219A CN 202111427220 A CN202111427220 A CN 202111427220A CN 113851219 A CN113851219 A CN 113851219A
- Authority
- CN
- China
- Prior art keywords
- knowledge
- graph
- entity
- patient
- entities
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
- G06F40/295—Named entity recognition
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/26—Speech to text systems
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
- G16H10/60—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for patient-specific data, e.g. for electronic patient records
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
Abstract
The invention provides an intelligent diagnosis guiding method based on a multi-mode knowledge graph, which comprises the steps of constructing the knowledge graph, obtaining a chief complaint text, a report single picture and a chief complaint audio input by a patient, making an entity list according to patient symptoms obtained by multi-information identification, and calculating by using a weight fusion algorithm to obtain a result of a department with rights. The invention relates to the field of knowledge maps in artificial intelligence, and the method utilizes a multi-mode knowledge map to conduct medical diagnosis guidance, comprises information identification of natural language, images and audio, and also conducts intelligent diagnosis guidance processing according to multi-dimensional information.
Description
Technical Field
The invention relates to the field of knowledge maps in artificial intelligence, in particular to a multi-mode knowledge map-based medical intelligent diagnosis guide.
Background
Because patients in a third-generation, large-scale and medium-scale hospital are increased rapidly, the manual diagnosis guiding workload is large, one-to-one service cannot be realized, the patients cannot find sickroom departments for examination and treatment quickly and accurately, and in addition, the phenomenon that a large number of patients queue and are tied up or 'mengkou' is rare due to the lack of timely understanding of queuing information, so that great challenge is brought to the diagnosis guiding work of the hospital. In order to improve the service quality and promote the harmony and stability of the doctor-patient relationship, a large amount of repeated diagnosis guide work needs to be completed by a mobile phone and a network, so that more manpower is put into more important work, limited medical resources are reasonably distributed and fully utilized, the medical service quality and the management level of the hospital are effectively improved, the complaints of patients can be reduced, and the harmony of the doctor-patient relationship is promoted.
The intelligent diagnosis guiding system is a system for recommending and registering consulting rooms according to information, examination sheets and voice input by users, some intelligent diagnosis guiding systems are proposed at home and abroad, traditional software technology is mostly adopted, people such as the beautiful jade apply for 'a multi-round conversation diagnosis guiding system and method', the multi-round conversation diagnosis guiding system and method are provided, the relation between symptoms and departments is established, and compared with other diagnosis guiding methods, the system is quicker and more accurate. The tsuga et al apply for an intelligent diagnosis guide processing method, device, electronic device and storage medium, along with the development of artificial intelligence, a diagnosis guide system based on a deep learning network is developed vigorously, if the later and the like apply for a semantic triage method and system, a long-short term memory network model is trained according to the historical inquiry data, and the department classification model is obtained.
However, the above schemes mostly use a single-source traditional knowledge graph, mainly focus on studying the entities and relations of texts and data, and it is difficult to well relate symptoms, diseases and departments together.
Disclosure of Invention
The invention aims to provide an efficient and intelligent diagnosis guide method based on the current single-source knowledge graph pre-diagnosis system, which couples a medical technology, a voice recognition technology and a video recognition technology and constructs a multi-mode knowledge graph.
In order to achieve the purpose, the invention is realized by the following technical scheme:
an intelligent diagnosis guiding method based on a multi-modal knowledge graph comprises the following steps:
(a) acquiring medical data to construct a knowledge graph;
(b) storing the knowledge graph into a graph database to form a knowledge graph database;
(c) acquiring a chief complaint text input by a patient, identifying and normalizing Chinese entities, namely recalling a normalized entity from a map node according to an identified entity result;
(d) acquiring a report picture uploaded by a patient, converting image data into character information, and recalling a standardized examination index name;
(e) acquiring a chief complaint audio input by a patient, performing voice recognition, and converting audio data into character data;
(f) and making an entity list according to the patient symptoms identified by the multivariate information, traversing the symptom graph to obtain the symptom-disease relation and the weight thereof, and calculating by using a weight fusion algorithm to obtain the result of the department with the right.
Further, the knowledge graph constructed by the medical data comprises a disease state knowledge graph constructed by medical text knowledge and a medical examination order knowledge graph, the disease state knowledge graph comprises the correlation of symptoms, diseases and departments, and the medical examination order knowledge graph comprises the corresponding relation of examination reports and the diseases.
Further, the medical checklist knowledge map also includes probabilistic relationships between individual symptoms and diseases.
And further, the step b comprises knowledge merging and knowledge storage, wherein the knowledge merging is to merge medical text knowledge and medical examination order knowledge, merge the matched entities through similarity calculation, and store the entities into a database to form a knowledge map database.
Further, step c includes:
(c1) denoising the main complaint text;
(c2) identifying the main complaint text input by the patient into three entities of symptoms, diseases and parts;
(c3) and recalling the normalized entity from the map node according to the entity result obtained by identification.
Further, the entity identification method comprises the following steps:
and constructing two dictionary trees by using nodes of symptoms and parts in the knowledge graph, and searching node data contained in the dictionary trees from the main complaint text to obtain an entity recognition result.
Further, the method for recalling the normalized entity comprises the following steps:
the part dictionary is used to extract the symptom entity and the object to be recalled, and entities with consistent parts are screened.
Further, the method for recalling the normalized entity further comprises the following steps:
and performing similarity comparison and replacement on the similar words obtained by identification.
Further, step f is:
according to the patient symptom entity list obtained by the multivariate information identification, traversing the pathology graph to obtain symptom-disease relation and weight thereof, calculating the weight of a certain disease suffered by the patient by using a weight fusion algorithm, according to the patient disease entity list obtained by the multivariate information identification, the report single abnormal index and the chief complaint audio frequency, updating the disease weight again, according to the disease-department relation fusion weight, obtaining a department result with a right, and finally recommending a department with higher weight to a user.
The invention has the advantages that:
the traditional knowledge graph mainly focuses on researching entities and relations of texts and databases, and the multi-modal knowledge graph constructs entities under multiple modes and multi-modal semantic relations among the entities of the multiple modes on the basis of the traditional knowledge graph.
According to the invention, the medical text knowledge and the medical examination order knowledge are fused according to the information of different modes to construct a knowledge map database, the medical text knowledge is more focused on spoken information used by a patient, the medical examination order knowledge is more focused on written and specialized characters used by a doctor, and the information contained by a multi-mode entity is better utilized to obtain a more accurate knowledge map.
The invention adopts the graph database to store the knowledge graph, can use the natural extension characteristic of the graph structure to design the query algorithm without index adjacent node traversal, namely the graph traversal algorithm design, can quickly and conveniently find out the adjacent nodes thereof, and quickens the identification speed.
The method adopts the operation of similar entity replacement in the process of entity normalization, so that words which have no difference in understanding but influence the accuracy of the character similarity algorithm due to the written difference do not generate interference any more, and the accuracy of entity linking is greatly improved.
The invention utilizes the multi-mode knowledge graph to conduct medical treatment guide treatment, comprises information identification of natural language, images and audio, and also conducts intelligent guide treatment according to multi-dimensional information, compared with the traditional knowledge graph, the obtained information is more comprehensive and perfect, and the obtained result is more accurate.
According to the invention, through two times of weight calculation, the weight calculation is carried out on the relation between symptoms and diseases, and then the pre-judged diseases, report sheet abnormal indexes and the voice input of the patient identified by the multivariate information are used for assisting, and the disease weight is updated again.
Drawings
Fig. 1 is a schematic flow chart of knowledge fusion and knowledge merging in embodiment 1 of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In this embodiment, an intelligent diagnosis guiding method based on a multi-modal knowledge graph is provided, which includes the steps of:
s1, constructing a pathology knowledge graph by using the existing medical text data: the medical data of the Wawei medicine website is crawled, and the screened map comprises the information of diseases and symptoms, and also comprises the correlation between symptoms and diseases and the correlation between diseases and departments.
S2, adopting the medical examination order knowledge to construct a medical examination order knowledge map: and crawling the diseases corresponding to the abnormal values of the conventional examination items to obtain the corresponding relation between the examination report and the diseases, and calculating to obtain the probability relation between the single symptom and the diseases.
And S3, performing knowledge fusion on the two knowledge maps, wherein the knowledge fusion is mainly characterized by entity matching and knowledge merging, and for the same symptom, entities corresponding to the medical text data are biased to be spoken, but entities corresponding to the examination order are biased to be written and specialized, so that the two preprocessed knowledge maps are required to be subjected to entity matching by using an entity matching method of similarity calculation. The specific steps can be seen in fig. 1, the data of the knowledge graph 1, namely the pathology knowledge graph, and the data of the knowledge graph 2, namely the medical examination order knowledge graph are preprocessed and then stored, the two knowledge graphs are preprocessed, namely missing values, repeated values and abnormal values in the data are processed, all the repeated values are deleted, the missing values are filled by using the last effective value, the abnormal values are judged by using a box line method, and the abnormal values are processed by using a 95-division capping method. In order to realize the expandability of data, a partitioning mechanism is adopted for access, the data is partitioned into different blocks and stored in different computers, the data of each block has two backup data, and when the machine is down due to network and other reasons, the data can be acquired from the backup blocks. And after data are obtained, attribute similarity calculation is carried out through a word meaning similarity algorithm-a synonym comparison algorithm based on WordNet, the similarity calculation of two aspects of the attribute similarity of the entity and the similarity of the obtained entity is linked to the entity operation in the knowledge base, and finally the same entity from different knowledge bases is matched to realize knowledge combination.
And S4, storing the fused knowledge graph into a graph database, for example, selecting a Neo4J graph database to store the knowledge graph. Neo4j is a native graph database engine that stores native graph data, and therefore can use the natural stretch property of graph structures to design query algorithms that are free of index neighbor traversal, i.e., graph traversal algorithm design. The traversal of the graph is a unique algorithm of the graph data structure, namely, starting from one node, and according to the connection relation, the adjacent nodes can be quickly and conveniently found out.
S5, obtaining the chief complaint text input by the patient, identifying and normalizing Chinese entities, namely identifying three entities of symptoms, diseases and parts, but because the disease description of the patient is different, entity normalization is required, and recalling a normalized entity from the knowledge map node according to the identified entity result. In the process of entity identification, patient complaints may contain too much irrelevant information to affect the result of entity identification, so that noise filtering is required before identification. In order to solve the problem that the entity accuracy rate of the model is low during short sentence recognition, two AC trees (Aho-Corasick automation) are constructed by using nodes of symptoms and parts in a map, node data contained in a dictionary tree is searched from an original sentence, and a final entity recognition result is obtained after the dictionary tree and the model result are subjected to deduplication.
And S6, acquiring the medical report data, identifying the data in the report uploaded by the patient, converting the data into characters, and analyzing abnormal data in the characters. In addition, since the names of the hospital examination indexes may be slightly different, an operation similar to entity normalization is required here, and a normalized examination index name is recalled.
S7, obtaining the voice input data of the patient, such as the audio answer in question-answering mode, converting the audio data of the patient into text data, and then performing the relevant operation of natural language processing, such as voice recognition by using vosk library.
S8, weight fusion: according to the patient symptom entity list obtained by multivariate information identification, traversing the symptom graph to obtain symptom-disease relation and weight thereof, and calculating the weight of a patient suffering from a certain disease by using a reasonable weight fusion algorithm; and identifying the patient pre-judging diseases, report sheet abnormal indexes and answers of the patient in a question and answer mode according to the multivariate information, updating the weight of the diseases again, fusing the weight according to the disease-department relationship to obtain a department result with the weight, and finally recommending departments with higher weight or displaying the department result with the weight to the user.
In another embodiment, the medical text data is obtained by using an open source technology "OpenKG" to construct a pathology knowledge map, the data source uses a crawler technology to obtain medical data crawled from a wareway medicine website, and the medical data have scientific and accuracy guaranteed to a certain extent and comprise 8764 diseases, 5998 symptoms, 54710 symptoms-disease correlation and 8806 diseases-department correlation after being screened.
The medical examination list knowledge graph adopts a CCKS2019 Chinese named entity recognition task to recognize the corresponding relation between an examination report and a disease, in addition, the probability relation between a single symptom and the disease needs to be supplemented, the joint search is carried out on the symptom-disease in a search engine according to the frequency calculation correlation degree of the symptom-disease pair in the search engine, the occurrence frequency of the symptom-disease pair on the internet is obtained as a numerator, and the denominator is the sum of all symptom joint search results of the disease. And crawling diseases corresponding to abnormal values of the conventional inspection items from the related data of the search engine, and writing the diseases into the JSON file.
In another embodiment, the Chinese entity normalization uses a Bert-LSTM-CRF Chinese entity recognition model based on LSTM-CRF pre-filtering noise, which is mainly composed of three parts: an LSTM-CRF entity identification model, a noise filtering algorithm and a Bert-LSTM-CRF entity identification model. The patient's chief complaint original sentence is firstly input into the LSTM-CRF model to obtain the BIO labeling list of the original sentence, and three types (symptoms, diseases and parts) of entities in the original sentence are found out according to the labeling list to form an entity list. And inputting the entity list and the original sentence of the patient's chief complaint into a noise filtering algorithm, wherein the noise filtering algorithm can filter out sentences which do not contain entities or are among parts and symptom entities according to the position and the entity type of the entities in the original sentence to obtain the original noise filtering sentence. And inputting the noise-filtering original sentence into a Bert-LSTM-CRF entity recognition model to obtain a BIO labeling list of the noise-filtering original sentence, thereby obtaining an entity list of the noise-filtering original sentence.
For example, "doctor is good, and in recent times, chest is often dull and painful, so that when the doctor wants to consult, the chest is often painful, and what medicine needs to be taken, how to relieve the symptom. The patient is not suffered from the heart disease, and the patient becomes chest pain and chest pain after noise filtering, and the patient is not suffered from the heart disease. "the final Chinese entity recognition result is: "chest", "pain", "chest", "pain" and "heart disease".
In another embodiment, special processing is done for Chinese entity normalization. And normalizing the symptom and disease entity sets identified by the entities into the knowledge graph by using a character string similarity algorithm Textdistance. In the normalization process, the entities and recalled objects are extracted by using the part dictionary, and entities with consistent parts are screened. For example, in the above case, although there is no difference in the Chinese comprehension, that is, "pain" and "pain" all describe the pain sensation of the chest region, since the difference in terms affects the accuracy of the word similarity algorithm, in the normalization, the word replacement operation is performed on the entity containing such words, for example, BM25 or Textdistance string similarity comparison algorithm is used. The BM25 is an algorithm for evaluating the correlation between search terms and documents, and is an algorithm proposed based on a probabilistic search model. Textdistance is a third-party library of python that uses the Levenstein distance, also known as the Levenshtein distance, which is a type of edit distance. The minimum number of editing operations required for converting one string into another string. The allowed editing operations include replacing one character with another, inserting one character, and deleting one character. Finally, "chest", "pain" and "heart disease" can all be replaced and combined as: the part- "chest", the symptom- "pain", the patient pre-judges the disease- "heart disease".
In another embodiment, since for natural scene images, the text position in the image is located first and then the recognition is performed, the recognition of the medical report data is decomposed into two steps of text detection and text recognition. Character detection adopts a CTPN algorithm, combines with CNN and LSTM deep networks, decouples characters distributed transversely, and combines small horizontal pieces into a text line by using rules. Character recognition adopts two end-to-end OCR technologies of CRNN OCR and attention OCR based on deep learning to recognize the detected text sequence with transversely indefinite length and extract the item name, the inspection result and the normal value range in the inspection list. And recalling the normalized project names by using the BM25 model, and finding out abnormal projects according to the comparison result values of the project names.
In another embodiment, the invention can rely on a mobile phone applet or an intelligent terminal and the like to conduct the diagnosis guide.
Finally, it should be noted that: the above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that are within the spirit and principle of the present invention are intended to be included in the scope of the present invention.
Claims (9)
1. An intelligent diagnosis guiding method based on a multi-modal knowledge graph is characterized by comprising the following steps:
(a) acquiring medical data to construct a knowledge graph;
(b) storing the knowledge graph into a graph database to form a knowledge graph database;
(c) acquiring a chief complaint text input by a patient, identifying and normalizing Chinese entities, namely recalling a normalized entity from a map node according to an identified entity result;
(d) acquiring a report picture uploaded by a patient, converting image data into character information, and recalling a standardized examination index name;
(e) acquiring a chief complaint audio input by a patient, performing voice recognition, and converting audio data into character data;
(f) and making an entity list according to the patient symptoms identified by the multivariate information, traversing the symptom graph to obtain the symptom-disease relation and the weight thereof, and calculating by using a weight fusion algorithm to obtain the result of the department with the right.
2. The multi-modal knowledge-graph-based intelligent diagnosis guide method according to claim 1, wherein the knowledge graph constructed by the medical data in the step a comprises a pathology knowledge graph constructed by medical text knowledge and a medical examination order knowledge graph, the pathology knowledge graph comprises the correlation of symptoms, diseases and departments, and the medical examination order knowledge graph comprises the corresponding relation of examination reports and diseases.
3. The multimodal knowledge base-guided intelligent diagnosis method according to claim 2, wherein the medical examination order knowledge base further comprises a probabilistic relationship between individual symptoms and diseases.
4. The multi-modal knowledge-graph-based intelligent diagnosis guiding method according to claim 2, wherein the step b comprises knowledge merging and knowledge storage, wherein the knowledge merging is to merge medical text knowledge and medical examination order knowledge, merge the matched entities through similarity calculation, and store the entities into a database to form a knowledge-graph database.
5. The multimodal knowledge-graph based intelligent referral method of claim 1 wherein step c comprises:
(c1) denoising the main complaint text;
(c2) identifying the main complaint text input by the patient into three entities of symptoms, diseases and parts;
(c3) and recalling the normalized entity from the map node according to the entity result obtained by identification.
6. The multi-modal knowledge-graph based intelligent approach to guided physicians as claimed in claim 5, wherein the entity identification method of step c2 comprises the steps of:
and constructing two dictionary trees by using nodes of symptoms and parts in the knowledge graph, and searching node data contained in the dictionary trees from the main complaint text to obtain an entity recognition result.
7. The multimodal knowledge-graph based intelligent diagnosis guiding method according to claim 5, wherein the method of recalling the normalized entities in step c3 comprises the steps of:
the part dictionary is used to extract the symptom entity and the object to be recalled, and entities with consistent parts are screened.
8. The multimodal knowledge-graph based intelligent referral method of claim 7 wherein the method of recalling the normalized entities of step c3 further comprises the steps of:
and performing similarity comparison and replacement on the similar words obtained by identification.
9. The multimodal knowledge-graph based intelligent referral method of claim 1 wherein step f is:
according to the patient symptom entity list obtained by the multivariate information identification, traversing the pathology graph to obtain symptom-disease relation and weight thereof, calculating the weight of a certain disease suffered by the patient by using a weight fusion algorithm, updating the disease weight again according to the patient disease entity list obtained by the multivariate information identification, the report single abnormal index and the chief complaint audio, and obtaining a department result with the weight and recommending the department result to a user according to the disease-department relation fusion weight.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111427220.9A CN113851219A (en) | 2021-11-29 | 2021-11-29 | Intelligent diagnosis guiding method based on multi-mode knowledge graph |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111427220.9A CN113851219A (en) | 2021-11-29 | 2021-11-29 | Intelligent diagnosis guiding method based on multi-mode knowledge graph |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113851219A true CN113851219A (en) | 2021-12-28 |
Family
ID=78982152
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111427220.9A Pending CN113851219A (en) | 2021-11-29 | 2021-11-29 | Intelligent diagnosis guiding method based on multi-mode knowledge graph |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113851219A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115274086A (en) * | 2022-09-27 | 2022-11-01 | 无码科技（杭州）有限公司 | Intelligent diagnosis guiding method and system |
| CN116340778A (en) * | 2023-05-25 | 2023-06-27 | 智慧眼科技股份有限公司 | Medical large model construction method based on multiple modes and related equipment thereof |
| CN117689963A (en) * | 2024-02-02 | 2024-03-12 | 南京邮电大学 | Visual entity linking method based on multi-mode pre-training model |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110085307A (en) * | 2019-04-04 | 2019-08-02 | 华东理工大学 | A kind of intelligent hospital guide's method and system based on the fusion of multi-source knowledge mapping |
| CN110489566A (en) * | 2019-08-22 | 2019-11-22 | 上海软中信息系统咨询有限公司 | A kind of hospital guide's method of intelligence hospital guide's service robot |
| CN111370102A (en) * | 2020-02-06 | 2020-07-03 | 清华大学 | Department diagnosis guiding method, device and equipment |
| CN112035674A (en) * | 2020-08-28 | 2020-12-04 | 康键信息技术（深圳）有限公司 | Diagnosis guide data acquisition method and device, computer equipment and storage medium |
| CN112786194A (en) * | 2021-01-28 | 2021-05-11 | 北京一脉阳光医学信息技术有限公司 | Medical image diagnosis guide inspection system, method and equipment based on artificial intelligence |
| CN113689951A (en) * | 2021-08-04 | 2021-11-23 | 翼健(上海)信息科技有限公司 | Intelligent diagnosis guiding method, system and computer readable storage medium |
-
2021
- 2021-11-29 CN CN202111427220.9A patent/CN113851219A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110085307A (en) * | 2019-04-04 | 2019-08-02 | 华东理工大学 | A kind of intelligent hospital guide's method and system based on the fusion of multi-source knowledge mapping |
| CN110489566A (en) * | 2019-08-22 | 2019-11-22 | 上海软中信息系统咨询有限公司 | A kind of hospital guide's method of intelligence hospital guide's service robot |
| CN111370102A (en) * | 2020-02-06 | 2020-07-03 | 清华大学 | Department diagnosis guiding method, device and equipment |
| CN112035674A (en) * | 2020-08-28 | 2020-12-04 | 康键信息技术（深圳）有限公司 | Diagnosis guide data acquisition method and device, computer equipment and storage medium |
| CN112786194A (en) * | 2021-01-28 | 2021-05-11 | 北京一脉阳光医学信息技术有限公司 | Medical image diagnosis guide inspection system, method and equipment based on artificial intelligence |
| CN113689951A (en) * | 2021-08-04 | 2021-11-23 | 翼健(上海)信息科技有限公司 | Intelligent diagnosis guiding method, system and computer readable storage medium |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115274086A (en) * | 2022-09-27 | 2022-11-01 | 无码科技（杭州）有限公司 | Intelligent diagnosis guiding method and system |
| CN116340778A (en) * | 2023-05-25 | 2023-06-27 | 智慧眼科技股份有限公司 | Medical large model construction method based on multiple modes and related equipment thereof |
| CN116340778B (en) * | 2023-05-25 | 2023-10-03 | 智慧眼科技股份有限公司 | Medical large model construction method based on multiple modes and related equipment thereof |
| CN117689963A (en) * | 2024-02-02 | 2024-03-12 | 南京邮电大学 | Visual entity linking method based on multi-mode pre-training model |
| CN117689963B (en) * | 2024-02-02 | 2024-04-09 | 南京邮电大学 | Visual entity linking method based on multi-mode pre-training model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106776711B (en) | Chinese medical knowledge map construction method based on deep learning | |
| CN112786194A (en) | Medical image diagnosis guide inspection system, method and equipment based on artificial intelligence | |
| US9621601B2 (en) | User collaboration for answer generation in question and answer system | |
| CN113851219A (en) | Intelligent diagnosis guiding method based on multi-mode knowledge graph | |
| US20170161619A1 (en) | Concept-Based Navigation | |
| CN111428053A (en) | Tax field knowledge graph construction method | |
| WO2021139232A1 (en) | Medical knowledge graph-based triage method and apparatus, device, and storage medium | |
| CN111897967A (en) | Medical inquiry recommendation method based on knowledge graph and social media | |
| US20150149461A1 (en) | System and method for analyzing unstructured data on applications, devices or networks | |
| US20170262783A1 (en) | Team Formation | |
| CN102663129A (en) | Medical field deep question and answer method and medical retrieval system | |
| KR102491172B1 (en) | Natural language question-answering system and learning method | |
| CN112614565A (en) | Traditional Chinese medicine classic famous prescription intelligent recommendation method based on knowledge-graph technology | |
| WO2020074023A1 (en) | Deep learning-based method and device for screening for key sentences in medical document | |
| WO2020010834A1 (en) | Faq question and answer library generalization method, apparatus, and device | |
| WO2023178971A1 (en) | Internet registration method, apparatus and device for seeking medical advice, and storage medium | |
| WO2023040493A1 (en) | Event detection | |
| CN113569023A (en) | Chinese medicine question-answering system and method based on knowledge graph | |
| CN110750540A (en) | Method for constructing medical service knowledge base, method and system for obtaining medical service semantic model and medium | |
| CN113764112A (en) | Online medical question and answer method | |
| CN113849657A (en) | Structured data processing method of intelligent supervision black box | |
| CN114911893A (en) | Method and system for automatically constructing knowledge base based on knowledge graph | |
| Xie | A tcm question and answer system based on medical records knowledge graph | |
| CN112632223B (en) | Case and event knowledge graph construction method and related equipment | |
| CN111460173B (en) | Method for constructing disease ontology model of thyroid cancer |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20211228 |
|
| RJ01 | Rejection of invention patent application after publication |