Traditional Chinese medicine evidence-based knowledge extraction method and system

Abstract

The invention provides a traditional Chinese medicine evidence-based knowledge extraction method and system. The method comprises the following steps: collecting diversified literature data related to traditional Chinese medicine from multiple channels; performing imaging processing on the diversified literature data to obtain a literature image, and performing chromatic aberration balance on the literature image to obtain a balanced image; performing image enhancement processing on the balanced image to obtain an enhanced image; performing OCR recognition according to the enhanced image to obtain a recognition text, and performing data cleaning on the recognition text to obtain to-be-detected data; and inputting to-be-tested data into the constructed multi-strategy knowledge extraction model to obtain an extraction result. The traditional Chinese medicine related literature data is collected through multiple channels, the universality and diversity of data sources are ensured, and therefore rich knowledge and research results in the traditional Chinese medicine field are reflected. In the processing process, the readability of the literature image is improved through the application of image processing and chromatic aberration balancing technologies, and high-quality input is provided for subsequent OCR recognition.

Description

Technical Field

[0001] The present invention relates to the technical field of knowledge extraction, and in particular to a method and system for extracting evidence-based knowledge of traditional Chinese medicine. Background Art

[0002] Evidence-based medicine (EBM), which means "medicine based on evidence", is also known as evidence-based medicine and is a medical diagnosis and treatment method that emphasizes the use of well-designed and well-executed research (evidence) to optimize decision-making.

[0003] Evidence-based medicine is an emerging interdisciplinary clinical medicine discipline developed in the early 1990s. It is a method based on clinical practice and the rational application of the best evidence. Evidence-based Chinese medicine draws on the theories and methods of evidence-based medicine to collect, evaluate, produce and transform evidence of the effectiveness, safety and economy of Chinese medicine. It can reveal the characteristics and laws of the clinical effects of Chinese medicine and guide the formulation of clinical guidelines, pathways and health decisions. It is an important branch of the discipline and research field of evidence-based medicine. In recent years, the number of clinical evidence-based research papers published in Chinese medicine has increased, but clinicians have rarely applied the diagnosis and treatment methods in the literature. On the one hand, the conclusions in the literature cannot solve complex problems in the clinic, and clinicians will give up using them because of the relatively single chain of evidence when looking for evidence for diagnosis and treatment decisions. On the other hand, clinical practice literature has not been directly converted into medical decision-making recommendations and cannot be consulted in real time and accurately, which has largely hindered the operability of evidence-based literature in clinical decision-making and practice.

[0004] The prior art with application number 202311159972.0 discloses a multi-strategy knowledge extraction system, which relates to the field of knowledge system technology. The key points of its technical solution are: the multi-strategy knowledge extraction system includes a data layer, a preprocessing layer, a model layer, a service layer and an application layer. In this knowledge extraction system, by constructing a multi-strategy knowledge extraction model, different extraction strategies can be formulated for documents of different categories, and a large amount of data can be mechanically annotated, and it is suitable for long texts and sparse words at the same time; based on the knowledge graph research framework, taking the clinical literature of evidence-based gynecology in traditional Chinese medicine as an example, a multi-dimensional multi-strategy knowledge extraction framework based on deep learning is constructed to realize the entity extraction of knowledge units in the field of evidence-based Chinese medicine, and to study the key technologies for constructing the evidence-based knowledge system of gynecology in traditional Chinese medicine, so as to provide the best clinical evidence for clinical research on gynecology in traditional Chinese medicine at home and abroad. However, there may be problems with the quality of literature and data sources, which affects the accuracy of knowledge extraction. In addition, the formats of different documents are diverse, which increases the complexity of preprocessing. Summary of the invention

[0005] In order to overcome the deficiencies of the prior art, the purpose of the present invention is to provide a method and system for extracting evidence-based knowledge of traditional Chinese medicine, which can effectively extract key knowledge units of traditional Chinese medicine and ensure the reliability and applicability of the knowledge.

[0006] To achieve the above object, the present invention provides the following solutions:

[0007] A method for extracting evidence-based knowledge of traditional Chinese medicine, comprising:

[0008] Collect diversified literature data related to traditional Chinese medicine from various channels; the diversified literature data include: academic journal articles, dissertations and research reports, books and monographs, conference papers and abstracts, clinical guidelines and consensus opinions, online databases, and government and industry reports;

[0009] Performing image processing on the diversified document data to obtain a document image, and performing color balance on the document image to obtain a balanced image;

[0010] Performing image enhancement processing on the balanced image to obtain an enhanced image;

[0011] Performing OCR recognition on the enhanced image to obtain recognized text, and performing data cleaning on the recognized text to obtain data to be tested;

[0012] The test data is input into the constructed multi-strategy knowledge extraction model to obtain the extraction result.

[0013] Preferably, the plurality of channels includes journals, books, databases and online resources.

[0014] Preferably, performing color balance on the document image to obtain a balanced image includes:

[0015] Calculate the pixel mean of the original document image in the three RGB channels; wherein the pixel mean calculation formula is:

[0016]

[0017] Where D represents the pixel mean, represents the pixel mean of the original document image in the red channel, represents the pixel mean of the original document image in the green channel, Represents the pixel mean of the original document image in the blue channel;

[0018] Calculating a balance coefficient for each color channel according to the average pixel value;

[0019] The color difference of the original document image is balanced by using the balance coefficient to obtain the balanced image.

[0020] Preferably, the formula using the balance coefficient is:

[0021]

[0022] Among them, K R Indicates the balance coefficient of the red channel, K G Indicates the balance coefficient of the green channel, K B represents the balance coefficient of the blue channel, R′ represents the pixel value of the red channel after balance, G′ represents the pixel value of the green channel after balance, B′ represents the pixel value of the blue channel after balance, R represents the pixel value of the original document image in the red channel, G represents the pixel value of the original document image in the green channel, and B represents the pixel value of the original document image in the blue channel.

[0023] Preferably, performing image enhancement processing on the balanced image to obtain an enhanced image includes:

[0024] Obtaining the average brightness and saturation of each pixel of the balanced image;

[0025] Calculate the brightness variance and saturation variance of each pixel point according to the brightness mean and saturation mean of each pixel point;

[0026] constructing an enhancement coefficient according to the brightness variance and the saturation variance;

[0027] constructing an enhancement model using the enhancement coefficient;

[0028] The enhanced model is used to perform image enhancement on the balanced image to obtain an enhanced image.

[0029] Preferably, the formula of the enhancement coefficient is:

[0030]

[0031] Wherein, W represents the number of pixels on the balanced image, υ(i,j) represents the brightness value at point (i,j), represents the mean brightness of all pixels in the balanced image, s(i,j) represents the saturation at point (i,j), and s W (x, y) represents the mean saturation of all pixels in the balanced image, σ υ (x, y) represents the brightness variance of all pixels in the balanced image, σ s (x, y) represents the saturation variance of all pixels in the balanced image.

[0032] Preferably, the data cleaning method includes: checking the consistency of input data, processing invalid data and missing data.

[0033] Preferably, the multi-strategy knowledge extraction model includes: a grammatical pattern sub-model, a domain dictionary rule sub-model and a deep learning sub-model.

[0034] Preferably, the extraction results include: conventional article concepts, evidence-based entities and entities in the field of traditional Chinese medicine.

[0035] A TCM evidence-based knowledge extraction system, comprising:

[0036] The collection module is used to collect diversified literature data related to traditional Chinese medicine from various channels;

[0037] A first image processing module, used for performing image processing on the diversified document data to obtain a document image, and performing color balance on the document image to obtain a balanced image;

[0038] A second image processing module, used for performing image enhancement processing on the balanced image to obtain an enhanced image;

[0039] A text acquisition module, used to perform OCR recognition according to the enhanced image to obtain recognized text, and perform data cleaning on the recognized text to obtain data to be tested;

[0040] The knowledge extraction module is used to input the test data into the constructed multi-strategy knowledge extraction model to obtain the extraction result.

[0041] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0042] The present invention provides a method and system for extracting evidence-based knowledge of traditional Chinese medicine, the method comprising: collecting diversified literature data related to traditional Chinese medicine from multiple channels; performing image processing on the diversified literature data to obtain a literature image, and performing color balance on the literature image to obtain a balanced image; performing image enhancement processing on the balanced image to obtain an enhanced image; performing OCR recognition based on the enhanced image to obtain a recognized text, and performing data cleaning on the recognized text to obtain data to be tested; inputting the data to be tested into a constructed multi-strategy knowledge extraction model to obtain an extraction result. The present invention collects literature data related to traditional Chinese medicine through multiple channels, ensuring the breadth and diversity of data sources, thereby reflecting the rich knowledge and research results in the field of traditional Chinese medicine. In the processing process, the application of image processing and color balance technology improves the readability of the literature image and provides high-quality input for subsequent OCR recognition. This series of image enhancement processing ensures the accuracy of the recognized text, reduces the impact of noise on data cleaning, and significantly improves the quality of the data to be tested, thereby effectively extracting key traditional Chinese medicine knowledge units and ensuring the reliability and applicability of knowledge. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0044] Figure 1 A flow chart of a method provided by an embodiment of the present invention;

[0045] Figure 2 A schematic diagram of the system structure provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] The purpose of the present invention is to provide a method and system for extracting evidence-based knowledge of traditional Chinese medicine, which can improve the quality of data to be tested.

[0048] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0049] Figure 1 A flow chart of a method provided by an embodiment of the present invention, such as Figure 1 As shown, the present invention provides a method for extracting evidence-based knowledge of traditional Chinese medicine, comprising:

[0050] Collect diversified literature data related to traditional Chinese medicine from various channels; the diversified literature data include: academic journal articles, dissertations and research reports, books and monographs, conference papers and abstracts, clinical guidelines and consensus opinions, online databases, and government and industry reports;

[0051] Performing image processing on the diversified document data to obtain a document image, and performing color balance on the document image to obtain a balanced image;

[0052] Performing image enhancement processing on the balanced image to obtain an enhanced image;

[0053] Performing OCR recognition on the enhanced image to obtain recognized text, and performing data cleaning on the recognized text to obtain data to be tested;

[0054] The test data is input into the constructed multi-strategy knowledge extraction model to obtain the extraction result.

[0055] Specifically, in the present invention, the diversified literature data includes academic journal articles, such as "Journal of Traditional Chinese Medicine" and "Chinese Journal of Traditional Chinese Medicine", which provide the latest research results and clinical trial data; master's and doctoral dissertations, especially those involving research in the field of traditional Chinese medicine, which provide in-depth literature reviews and experimental data; classic works of traditional Chinese medicine, modern traditional Chinese medicine textbooks and professional books, covering theoretical knowledge, treatment methods and case analysis; academic conference papers and abstracts related to traditional Chinese medicine, which show the latest research trends and academic discussions; clinical practice guidelines and expert consensus, which provide clinical application suggestions for evidence-based medicine; online databases and resources, such as China National Knowledge Infrastructure (CNKI), Wanfang Data and PubMed, which provide rich literature resources; government and industry reports, which publish development reports, policy documents and statistical data of traditional Chinese medicine; and real patient cases and clinical records, which provide application examples and effect evaluation in practice. These diversified literature data provide a comprehensive basis for knowledge extraction and reflect the research status and development trends in the field of traditional Chinese medicine.

[0056] Preferably, the plurality of channels includes journals, books, databases and online resources.

[0057] Further, in the present invention, the channels for collecting diversified literature data related to traditional Chinese medicine include academic journals and databases, such as China National Knowledge Infrastructure (CNKI), Wanfang Data, and PubMed, which provide a wide range of research papers and journal articles; libraries of universities and research institutions, which obtain dissertations and research reports; academic conferences related to traditional Chinese medicine, which obtain conference papers and abstracts; reports and policy documents issued by governments and industry associations, which provide authoritative industry data and development trends; online bookstores and publishers, which obtain traditional Chinese medicine classics and modern teaching materials; and hospitals and clinical research centers, which collect real patient cases and clinical records. These channels ensure the diversity and comprehensiveness of data sources and provide a rich foundation for the extraction of evidence-based knowledge of traditional Chinese medicine.

[0058] Preferably, performing color balance on the document image to obtain a balanced image includes:

[0059] Calculate the pixel mean of the original document image in the three RGB channels; wherein the pixel mean calculation formula is:

[0060]

[0061] Where D represents the pixel mean, represents the pixel mean of the original document image in the red channel, represents the pixel mean of the original document image in the green channel, Represents the pixel mean of the original document image in the blue channel;

[0062] Calculating a balance coefficient for each color channel according to the average pixel value;

[0063] The color difference of the original document image is balanced by using the balance coefficient to obtain the balanced image.

[0064] Preferably, the formula using the balance coefficient is:

[0065]

[0066] Among them, K R Indicates the balance coefficient of the red channel, K G Indicates the balance coefficient of the green channel, K B represents the balance coefficient of the blue channel, R′ represents the pixel value of the red channel after balance, G′ represents the pixel value of the green channel after balance, B′ represents the pixel value of the blue channel after balance, R represents the pixel value of the original document image in the red channel, G represents the pixel value of the original document image in the green channel, and B represents the pixel value of the original document image in the blue channel.

[0067] Furthermore, the present invention takes into account that a considerable portion of the numerous periodical materials are in the form of pictures. When the external lighting conditions change, the surface color of the book paper remains unchanged, but the color of the object collected by the imaging device will deviate with the change of lighting, gradually deviating from the true color of the book paper. The present invention uses a color difference balance algorithm to compensate for this defect of the imaging device and correct the color difference of the image.

[0068] Preferably, performing image enhancement processing on the balanced image to obtain an enhanced image includes:

[0069] Obtaining the average brightness and saturation of each pixel of the balanced image;

[0070] Calculate the brightness variance and saturation variance of each pixel point according to the brightness mean and saturation mean of each pixel point;

[0071] constructing an enhancement coefficient according to the brightness variance and the saturation variance;

[0072] constructing an enhancement model using the enhancement coefficient;

[0073] The enhanced model is used to perform image enhancement on the balanced image to obtain an enhanced image.

[0074] Preferably, the formula of the enhancement coefficient is:

[0075]

[0076] Wherein, W represents the number of pixels on the balanced image, υ(i,j) represents the brightness value at point (i,j), represents the mean brightness of all pixels in the balanced image, s(i,j) represents the saturation at point (i,j), and s W (x, y) represents the mean saturation of all pixels in the balanced image, σ υ (x, y) represents the brightness variance of all pixels in the balanced image, σ s (x, y) represents the saturation variance of all pixels in the balanced image.

[0077] Preferably, the data cleaning method includes: checking the consistency of input data, processing invalid data and missing data.

[0078] Specifically, the data cleaning process of this embodiment includes the following steps: first, check the consistency of input data, and ensure that all data conforms to the expected structure and type by defining standard formats and rules; then, process invalid data, and ensure the accuracy and relevance of data by identifying and deleting duplicate, erroneous or irrelevant data entries; finally, for missing data, use interpolation, mean filling or context-based inference techniques to fill in, so as to reduce the impact of incomplete data on analysis results. Demonstration, this embodiment is implemented through automated scripts and data processing tools to ensure the efficiency and reliability of the data cleaning process and provide high-quality input data for subsequent knowledge extraction.

[0079] Preferably, the multi-strategy knowledge extraction model includes: a grammatical pattern sub-model, a domain dictionary rule sub-model and a deep learning sub-model.

[0080] Specifically, in the multi-strategy knowledge extraction model, the application process of each sub-model works together. First, the grammatical pattern sub-model extracts preliminary knowledge fragments by parsing the syntactic structure in the text and identifying specific language patterns, such as TCM terms and conceptual relationships. The domain dictionary rule sub-model uses pre-defined TCM terminology dictionaries and rules to scan texts through pattern matching, identify and annotate key terms and their contexts, and improve the accuracy of knowledge extraction and applicability to specific fields. Finally, the deep learning sub-model is trained with a large amount of annotated data, and uses a neural network structure to automatically learn complex semantic and contextual relationships, enhancing the ability to recognize unseen data and diverse expressions. The three sub-models are organically combined to achieve multi-level and multi-perspective knowledge extraction, improving the comprehensiveness and reliability of the extraction.

[0081] Preferably, the extraction results include: conventional article concepts, evidence-based entities and entities in the field of traditional Chinese medicine.

[0082] Corresponding to the above method, such as Figure 2 As shown, this embodiment also provides a TCM evidence-based knowledge extraction system, including:

[0083] The collection module is used to collect diversified literature data related to traditional Chinese medicine from various channels;

[0084] A first image processing module, used for performing image processing on the diversified document data to obtain a document image, and performing color balance on the document image to obtain a balanced image;

[0085] A second image processing module, used for performing image enhancement processing on the balanced image to obtain an enhanced image;

[0086] A text acquisition module, used to perform OCR recognition according to the enhanced image to obtain recognized text, and perform data cleaning on the recognized text to obtain data to be tested;

[0087] The knowledge extraction module is used to input the test data into the constructed multi-strategy knowledge extraction model to obtain the extraction result.

[0088] The present invention has significant beneficial effects, especially in improving data quality and diversity. By collecting TCM-related literature data through multiple channels, the extensiveness and diversity of data sources are ensured, thereby reflecting the rich knowledge and research results in the field of TCM. During the processing, the application of image processing and color difference balance technology improves the readability of document images and provides high-quality input for subsequent OCR recognition. This series of image enhancement processing ensures the accuracy of text recognition, reduces the impact of noise on data cleaning, and significantly improves the quality of the data to be tested. Finally, through the multi-strategy knowledge extraction of the deep learning model, key TCM knowledge units can be effectively extracted, ensuring the reliability and applicability of the knowledge. The present invention not only improves the quality of the data, but also provides a solid foundation for evidence-based research in the field of TCM, and promotes the in-depth development of related scientific research work.

[0089] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0090] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for extracting evidence-based knowledge of traditional Chinese medicine, characterized in that: include: Collect diversified literature data related to traditional Chinese medicine from various channels; The diverse literature data includes: academic journal articles, dissertations and research reports, books and monographs, conference papers and abstracts, clinical guidelines and consensus opinions, online databases, and government and industry reports; Performing image processing on the diversified document data to obtain a document image, and performing color balance on the document image to obtain a balanced image; Performing image enhancement processing on the balanced image to obtain an enhanced image; Performing OCR recognition on the enhanced image to obtain recognized text, and performing data cleaning on the recognized text to obtain data to be tested; The test data is input into the constructed multi-strategy knowledge extraction model to obtain the extraction result.

2. The method for extracting evidence-based knowledge of traditional Chinese medicine according to claim 1, characterized in that: The multiple channels include journals, books, databases and online resources.

3. The method for extracting evidence-based knowledge of traditional Chinese medicine according to claim 1, characterized in that: Performing color difference balance on the document image to obtain a balanced image includes: Calculate the pixel mean of the original document image in the three RGB channels; wherein the pixel mean calculation formula is: Where D represents the pixel mean, represents the pixel mean of the original document image in the red channel, G represents the pixel mean of the original document image in the green channel, Represents the pixel mean of the original document image in the blue channel; Calculating a balance coefficient for each color channel according to the average pixel value; The color difference of the original document image is balanced by using the balance coefficient to obtain the balanced image.

4. The method for extracting evidence-based knowledge of traditional Chinese medicine according to claim 3, characterized in that: The formula using the balance coefficient is: Among them, K R Indicates the balance coefficient of the red channel, K G Indicates the balance coefficient of the green channel, K B represents the balance coefficient of the blue channel, R′ represents the pixel value of the red channel after balance, G′ represents the pixel value of the green channel after balance, B′ represents the pixel value of the blue channel after balance, R represents the pixel value of the original document image in the red channel, G represents the pixel value of the original document image in the green channel, and B represents the pixel value of the original document image in the blue channel.

5. The method for extracting evidence-based knowledge of traditional Chinese medicine according to claim 4, characterized in that: Performing image enhancement processing on the balanced image to obtain an enhanced image includes: Obtaining the average brightness and saturation of each pixel of the balanced image; Calculate the brightness variance and saturation variance of each pixel point according to the brightness mean and saturation mean of each pixel point; constructing an enhancement coefficient according to the brightness variance and the saturation variance; constructing an enhancement model using the enhancement coefficient; The enhanced model is used to perform image enhancement on the balanced image to obtain an enhanced image.

6. The method for extracting evidence-based knowledge of traditional Chinese medicine according to claim 5, characterized in that: The formula for the enhancement factor is: Wherein, W represents the number of pixels on the balanced image, υ(i,j) represents the brightness value at point (i,j), represents the mean brightness of all pixels in the balanced image, s(i,j) represents the saturation at point (i,j), and s W (x, y) represents the mean saturation of all pixels in the balanced image, σ υ (x, y) represents the brightness variance of all pixels in the balanced image, σ s (x, y) represents the saturation variance of all pixels in the balanced image.

7. The method for extracting evidence-based knowledge of traditional Chinese medicine according to claim 1, characterized in that: The data cleaning method includes: checking the consistency of input data, processing invalid data and missing data.

8. The method for extracting evidence-based knowledge of traditional Chinese medicine according to claim 1, characterized in that: The multi-strategy knowledge extraction model includes: a grammatical pattern sub-model, a domain dictionary rule sub-model and a deep learning sub-model.

9. The method for extracting evidence-based knowledge of traditional Chinese medicine according to claim 8, characterized in that: The extraction results include: conventional article concepts, evidence-based entities and entities in the field of traditional Chinese medicine.

10. A TCM evidence-based knowledge extraction system, characterized in that: include: The collection module is used to collect diversified literature data related to traditional Chinese medicine from various channels; The diverse literature data includes: academic journal articles, dissertations and research reports, books and monographs, conference papers and abstracts, clinical guidelines and consensus opinions, online databases, and government and industry reports; A first image processing module, used for performing image processing on the diversified document data to obtain a document image, and performing color balance on the document image to obtain a balanced image; A second image processing module, used for performing image enhancement processing on the balanced image to obtain an enhanced image; A text acquisition module, used to perform OCR recognition according to the enhanced image to obtain recognized text, and perform data cleaning on the recognized text to obtain data to be tested; The knowledge extraction module is used to input the test data into the constructed multi-strategy knowledge extraction model to obtain the extraction result.

Images