Blood pathology image analysis system based on artificial intelligence

Abstract

The invention discloses a blood pathology image analysis system based on artificial intelligence, and relates to the technical field of pathology image analysis, the blood pathology image analysis system comprises a pathology image database module, a sample experiment module, a pathology analysis module, a similar pathology retrieval module, a treatment course segmentation module, a medicine medical analysis module and a stage medical module, the system has the advantages that pathological images can be searched and classified more quickly through the pathological image database module, reference and comparison are provided for analyzing new pathological samples, image information with the highest similarity can be retrieved through the similar pathological retrieval module, and treatment data are extracted to generate treatment course information; the treatment experience of similar cases can be quickly referred, the medicine medical analysis module is beneficial to reasonably selecting medicines and evaluating the medicine treatment effect, the treatment scheme can be dynamically adjusted according to the actual condition of a patient through the stage medical module, the treatment scheme can be adjusted according to different conditions, and the flexibility of the scheme is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of pathological image analysis, and in particular to a blood pathological image analysis system based on artificial intelligence. Background Art

[0002] Hematopathology is an important branch of pathology, which mainly studies the pathological changes of blood system diseases. The blood system includes hematopoietic tissues such as red blood cells, white blood cells, platelets and bone marrow. Hematopathology diagnoses various blood system diseases, such as leukemia, lymphoma, myelodysplastic syndrome, etc., by carefully observing and analyzing the morphology, structure and quantity of cells in the blood and bone marrow. In hematopathology examination, bone marrow puncture biopsy and peripheral blood smear examination are usually performed. After obtaining the sample, the morphological characteristics of the cells are observed under a microscope, including abnormal changes in cell size, nuclear-cytoplasmic ratio, nuclear morphology, nuclear chromatin, cytoplasmic color and granules. These abnormal changes are often an important basis for judging the type and severity of blood system diseases.

[0003] Common blood pathology image analysis systems are difficult to accurately retrieve completely matching similar image information when in use, and cannot fully meet the treatment needs of complex diseases. They lack in-depth consideration of individual differences among patients, resulting in treatment plans that cannot adapt well to the patient's specific condition and economic situation. To this end, we propose a blood pathology image analysis system based on artificial intelligence. Summary of the invention

[0004] The purpose of the present invention is to provide a blood pathology image analysis system based on artificial intelligence.

[0005] In order to solve the problems raised in the above background technology, the present invention provides the following technical solutions: a blood pathology image analysis system based on artificial intelligence, including a patient information acquisition module, the patient information acquisition module is used to obtain the patient's pathological information, collect pathological samples, obtain the patient's blood data, and analyze the patient's disease information according to the blood data, and also includes a pathological image database module, a sample experiment module, a pathological analysis module, a similar pathology retrieval module and a treatment course segmentation module;

[0006] The pathological image database module can collect the previous pathological image information, generate pathological lesion images, establish a pathological image database, record the pathological lesion image information in the pathological image database, analyze the difference data between the pathological lesion image information and the normal pathological image information, generate a pathological difference table, mark the pathological difference table on the corresponding pathological image information, and determine the type of disease to which the pathological lesion image information belongs, and establish a disease folder according to the disease type;

[0007] The sample experiment module extracts the patient's pathological samples and then conducts experimental tests on the pathological samples;

[0008] The pathological analysis module analyzes whether the pathological sample has lesions. If the pathological sample has lesions, the pathological sample information is analyzed to determine the type of disease. The pathological sample information is then transmitted to the similar pathological retrieval module to generate abnormal sample information. If the pathological sample has no lesions, the pathological sample is marked with a normal label.

[0009] A similar pathology retrieval module collects abnormal sample information and extracts pathological lesion image information, compares the abnormal sample information with the pathological lesion image information, retrieves the pathological lesion image information with the highest similarity to the abnormal sample information, and then extracts the treatment data of the pathological lesion image information to generate a layer of treatment course information;

[0010] The treatment course segmentation module extracts the type of disease and analyzes the medical plan for the type of disease based on the type of disease and the first-level treatment course information, thereby generating second-level treatment course information and analyzing its treatment unit. The second-level treatment plan is then separated according to the treatment course unit to generate sub-treatment plans.

[0011] As a further solution of the present invention: it also includes a drug medical analysis module and a stage medical module;

[0012] The drug medical analysis module collects drugs for treating different types of conditions, analyzes the effects of different drugs in different sub-treatment plans, and generates medical effect data;

[0013] The stage medical treatment module extracts medical plan data, further integrates sub-treatment plans, generates sub-stage plans, combines different sub-stage plans into stage treatment plans, establishes interval time analysis units and past efficacy thresholds, and uses interval time analysis units and past efficacy thresholds to analyze sub-stage plans.

[0014] As a further solution of the present invention: when generating the pathological difference table, the pathological image database module will synchronously analyze the pathological features of different conditions and generate original pathological features. After analyzing the original pathological features of different conditions, the analyzed original pathological features will be recorded in the condition folder, and a weight adding function will be set in the condition folder. Then, the pathological features of the pathological lesion image information will be extracted to generate the current pathological features. The weight adding function information in the condition folder is set to B1, B2, B3, ..., B W , assuming the current pathological features are B1, B2, B3, ..., B T .

[0015] when When , it can be directly judged that the pathological lesion image of the pathological feature belongs to the disease type corresponding to the weighted addition function information in the disease folder, and no signal will be issued;

[0016] when When , it can be directly judged that the pathological lesion image of the pathological feature does not belong to the disease type corresponding to the weighted addition function information in the disease folder, and a feature matching signal will be issued;

[0017] According to the above method, the type of disease to which the pathological lesion image belongs can be preliminarily determined.

[0018] As a further solution of the present invention: after the pathological image database module preliminarily determines the type of disease to which the pathological lesion image belongs, it will receive a signal sent when the pathological lesion image is preliminarily determined. After receiving the feature matching signal, it will match the type of disease to which the pathological lesion image belongs through the pathological features. Suppose the original pathological features are Y1, Y2, Y3, ..., Y Z , let the feature matching degree be T 匹配度 , let the summary formula of subset quantity be

[0019]

[0020] The feature matching degree of the pathological lesion image can be calculated according to the above formula, and then the corresponding pathological lesion image is recorded in the disease folder corresponding to the original pathological feature with the highest feature matching degree.

[0021] As a further solution of the present invention: when calculating the feature matching degree of the pathological lesion image, the situation of the same feature matching degree may occur. At this time, a folder with the same matching degree is established, and then the pathological lesion image information with the same feature matching degree is recorded in the folder with the same matching degree, and the disease condition folder with the same feature matching degree is extracted. When the pathological lesion image information exists in the folder with the same matching degree, a judgment signal is transmitted to the doctor, the doctor retrieves the content in the folder with the same matching degree, and then the doctor makes a judgment on the disease condition folder to which the pathological lesion image information belongs.

[0022] As a further solution of the present invention: after determining that a pathological sample has a lesion, the pathological analysis module will transmit the pathological sample information to the pathological image database module, use the pathological image database module to determine the type of disease condition of the pathological sample information, and simultaneously convert the pathological sample information into pathological lesion image information and record it in the corresponding disease condition folder.

[0023] As a further solution of the present invention: the similar pathology retrieval module can receive information processed by the pathology analysis module, and when judging the type of the pathological sample information, it will simultaneously extract the pathological lesion image information in the disease type, and split the pathological sample information and the pathological lesion image information into diagnostic components, generate a sample component table and a lesion component table, and then calculate the similarity between the sample component table and the lesion component table. Suppose the component data in the sample component table is Assume that the component data in the lesion component table is Assume that the number of components in the sample component table and the lesion component table is M, and the similarity value is X 相似度 ;

[0024]

[0025] According to the above formula, the similarity value between the sample component table and the lesion component table can be calculated.

[0026] As a further solution of the present invention: when generating a sub-treatment plan, the treatment course segmentation module will establish a unit connection function and set default permissions for the unit connection function. Before generating the sub-treatment plan, the treatment course segmentation module will connect different treatment units through the unit connection function and merge them into a treatment unit again. At the same time, the doctor edits the information in the unit connection function so that it can be adjusted according to the doctor's cognition.

[0027] As a further solution of the present invention: when the stage medical module is combined into a stage treatment plan, it will use the interval time unit and the past efficacy threshold to analyze the maximum time that can be interval between different sub-stage plans, and at the same time grant the patient editing authority of the interval time unit. When the maximum time that can be interval between different sub-stage plans is analyzed, the interval time unit will be restricted according to the maximum time that can be interval between different sub-stage plans. The past efficacy threshold refers to the threshold at which the current sub-stage plan will not affect the treatment effect of the next sub-stage plan after the treatment is completed.

[0028] By adopting the above technical solution, compared with the prior art, the beneficial effects of the present invention are:

[0029] 1. The present invention can more quickly search and classify pathological images through the pathological image database module, provide reference and comparison for analyzing new pathological samples, use the similar pathological retrieval module to retrieve the image information with the highest similarity and extract the treatment data to generate a layer of treatment course information, which can quickly draw on the treatment experience of similar cases and improve the treatment efficiency. The treatment course segmentation module can separate the treatment units to generate sub-treatment plans, making the treatment plans more detailed and targeted. The drug medical analysis module helps to reasonably select drugs and evaluate the drug treatment effects. The stage medical module can dynamically adjust the treatment plan according to the actual situation of the patient, so that it can be adjusted according to different conditions, improving the flexibility of the plan;

[0030] 2. The present invention can deeply analyze the unique pathological characteristics of different conditions through the pathological image database module, providing a key basis for the subsequent accurate judgment of the type of condition to which the pathological lesion image belongs, and can make different conditions have a more refined weight distinction in the database, improve the accuracy and efficiency of judgment, and at the same time can accurately quantify the degree of fit between the pathological lesion image and the original pathological characteristics, making the classification and storage of pathological images more scientific and reasonable. Through the matching folder, the professional experience and subjective judgment ability of the doctor can be fully utilized to ensure that the pathological lesion image can finally be accurately classified into the appropriate condition folder, laying a solid and reliable foundation for subsequent medical analysis and treatment plan formulation;

[0031] 3. The present invention can realize the effective linkage between pathological sample information and pathological image database through the pathological analysis module, ensure the accuracy and consistency of disease type judgment, and provide reliable disease condition basis for subsequent treatment. The similar pathology retrieval module can provide strong support for determining appropriate treatment direction and reference. The treatment course segmentation module can flexibly adjust the combination of treatment units according to actual conditions, so that the sub-a treatment plan is more in line with the individual differences of patients. The stage medical treatment module not only ensures the continuity and effectiveness of the treatment process, but also fully considers the actual situation and treatment effect of the patient, making the formulation of the stage treatment plan more scientific and reasonable. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 Schematic diagram of the system flow in an embodiment of the present invention. DETAILED DESCRIPTION

[0033] The specific embodiments of the present invention will be further described below in conjunction with the accompanying drawings. It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention.

[0034] In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0035] Embodiment 1:

[0036] Therefore, in order to effectively solve the above problems, the present application proposes a blood pathology image analysis system based on artificial intelligence, as shown in the accompanying drawings of the specification. Figure 1 As shown, it includes a patient information acquisition module, which is used to obtain the patient's pathological information, collect pathological samples, obtain the patient's blood data, and analyze the patient's disease information based on the blood data. It also includes a pathological image database module, a sample experiment module, a pathological analysis module, a similar pathology retrieval module, and a treatment course segmentation module;

[0037] The pathological image database module can collect the previous pathological image information, generate pathological lesion images, establish a pathological image database, record the pathological lesion image information in the pathological image database, analyze the difference data between the pathological lesion image information and the normal pathological image information, generate a pathological difference table, mark the pathological difference table on the corresponding pathological image information, and determine the type of disease to which the pathological lesion image information belongs, and establish a disease folder according to the disease type;

[0038] Regularly reanalyze a large amount of pathological image information in the database, and automatically adjust the weights of different conditions in the pathological difference table according to newly emerging pathological features and types of conditions, so that the difference table can more accurately reflect the characteristics and differences of various conditions in current clinical practice;

[0039] The sample experiment module extracts the patient's pathological samples and then conducts experimental tests on the pathological samples;

[0040] The pathological analysis module analyzes whether the pathological sample has lesions. If the pathological sample has lesions, the pathological sample information is analyzed to determine the type of disease. The pathological sample information is then transmitted to the similar pathological retrieval module to generate abnormal sample information. If the pathological sample has no lesions, the pathological sample is marked with a normal label.

[0041] A similar pathology retrieval module collects abnormal sample information and extracts pathological lesion image information, compares the abnormal sample information with the pathological lesion image information, retrieves the pathological lesion image information with the highest similarity to the abnormal sample information, and then extracts the treatment data of the pathological lesion image information to generate a layer of treatment course information;

[0042] Combine the clinical symptom description of pathological samples, laboratory test indicators and other modal data with abnormal sample information for comprehensive retrieval, so as to retrieve the pathological lesion image information with the highest similarity more comprehensively and accurately, and improve the reliability and effectiveness of the retrieval;

[0043] The treatment course segmentation module extracts the disease type and analyzes the medical plan of the disease type according to the disease type and the first-level treatment course information, thereby generating the second-level treatment course information and analyzing its treatment unit. Then, the second-level treatment plan is separated according to the treatment course unit to generate a sub-treatment plan.

[0044] It also includes a drug medical analysis module and a stage medical module;

[0045] The drug medical analysis module collects drugs for treating different types of conditions, analyzes the effects of different drugs in different sub-treatment plans, and generates medical effect data;

[0046] The phase medical treatment module extracts medical treatment plan data, further integrates sub-treatment plans, generates sub-phase plans, combines different sub-phase plans into phase treatment plans, establishes interval time analysis units and past efficacy thresholds, and uses the interval time analysis units and past efficacy thresholds to analyze sub-phase plans;

[0047] Specific workflow: obtaining the patient's pathological information, collecting pathological samples, obtaining the patient's blood data, analyzing the patient's disease information based on the blood data, collecting previous pathological image information, generating pathological lesion images, analyzing the difference data between the pathological lesion image information and the normal pathological image information, generating a pathological difference table, marking the pathological difference table on the corresponding pathological image information, determining the type of disease to which the pathological lesion image information belongs, conducting experimental tests on the pathological samples, analyzing whether the pathological samples have lesions, generating abnormal sample information, extracting pathological lesion image information, retrieving the pathological lesion image information with the highest similarity to the abnormal sample information, generating a single course of treatment information, analyzing the medical plan for the type of disease, generating a double course of treatment information, analyzing its treatment unit, and then separating the double treatment plan according to the course of treatment unit to generate sub-treatment plans, collecting drugs for treating the type of disease, analyzing the effects of different drugs in different sub-treatment plans, integrating the sub-treatment plans, generating sub-stage plans, and analyzing the sub-stage plans using the interval time analysis unit and the past efficacy threshold;

[0048] Furthermore, the pathology image database module can more quickly search and classify pathology images, providing reference and comparison for analyzing new pathology samples. The similar pathology retrieval module can retrieve the image information with the highest similarity and extract treatment data to generate a layer of treatment course information, which can quickly draw on the treatment experience of similar cases and improve treatment efficiency. The treatment course segmentation module can separate the treatment units to generate sub-treatment plans, making the treatment plans more detailed and targeted. The drug medical analysis module helps to rationally select drugs and evaluate the effects of drug treatment. The stage medical module can dynamically adjust the treatment plan according to the actual situation of the patient, so that it can be adjusted according to different conditions, thereby improving the flexibility of the plan.

[0049] Embodiment 2:

[0050] Based on the first embodiment, as shown in the accompanying drawings of the specification Figure 1 As shown in FIG. 1 , when the pathological image database module generates the pathological difference table, it will simultaneously analyze the pathological characteristics of different conditions and generate the original pathological characteristics. After analyzing the original pathological characteristics of different conditions, it will record the analyzed original pathological characteristics in the condition folder, set the weight adding function in the condition folder, and then extract the pathological characteristics of the pathological lesion image information to generate the current pathological characteristics. Suppose the weight adding function information in the condition folder is B1, B2, B3, …, BW , assuming the current pathological features are B1, B2, B3, ..., B T .

[0051] when When , it can be directly judged that the pathological lesion image of the pathological feature belongs to the disease type corresponding to the weighted addition function information in the disease folder, and no signal will be issued;

[0052] when When , it can be directly judged that the pathological lesion image of the pathological feature does not belong to the disease type corresponding to the weighted addition function information in the disease folder, and a feature matching signal will be issued;

[0053] According to the above method, the type of disease to which the pathological lesion image belongs can be preliminarily determined;

[0054] Set up an automatic weight adjustment mechanism. As new pathological cases are continuously added to the database, the system will automatically adjust the weight of the pathological feature in the disease folder according to the frequency of occurrence of the pathological feature in these new cases and its importance to the disease diagnosis, so that the weight can reflect the value and significance of the pathological feature in clinical practice in real time;

[0055] After the pathological image database module preliminarily determines the type of disease to which the pathological lesion image belongs, it will receive the signal sent when the pathological lesion image is preliminarily determined. After receiving the feature matching signal, it will match the type of disease to which the pathological lesion image belongs through the pathological features. Suppose the original pathological features are Y1, Y2, Y3, ..., Y Z , let the feature matching degree be T 匹配度 , let the summary formula of subset quantity be

[0056]

[0057] According to the above formula, the feature matching degree of the pathological lesion image can be calculated, and then the corresponding pathological lesion image is recorded in the disease folder corresponding to the original pathological feature with the highest feature matching degree;

[0058] Combine the texture features, color features, morphological features and other multi-dimensional information of the pathological image, and perform a comprehensive matching calculation with the original pathological features, so as to determine the feature matching degree more comprehensively and scientifically, and reduce the error that may be caused by a single factor;

[0059] When calculating the feature matching degree of the pathological lesion image, there will be a situation where the feature matching degree is the same. At this time, a folder with the same matching degree is established, and then the pathological lesion image information with the same feature matching degree is recorded in the folder with the same matching degree, and the disease condition folder with the same feature matching degree is extracted. When the pathological lesion image information exists in the folder with the same matching degree, a judgment signal is transmitted to the doctor, and the doctor retrieves the content in the folder with the same matching degree, and then the doctor judges the disease condition folder to which the pathological lesion image information belongs;

[0060] Establish a pathological feature evolution recording module to record the changes of the pathological features in different time periods, as well as the relationship between these changes and the treatment process and disease progression, to provide important data support for subsequent research on the dynamic changes of pathological features and the evaluation of treatment effects;

[0061] Specific workflow: Analyze the pathological features of different conditions, generate original pathological features, record the analyzed original pathological features into the condition folder, set the weight adding function, extract the pathological features of the pathological lesion image information, preliminarily judge the type of condition to which the pathological lesion image belongs, receive the signal sent when preliminarily judging the type of condition to which the pathological lesion image belongs, match the type of condition to which the pathological lesion image belongs through the pathological features, calculate the feature matching degree of the pathological lesion image, record the corresponding pathological lesion image into the condition folder corresponding to the original pathological feature with the highest feature matching degree, establish a folder with the same matching degree, record the pathological lesion image information with the same feature matching degree into the folder with the same matching degree, transmit the judgment signal to the doctor, and the doctor will judge the condition folder to which the pathological lesion image information belongs;

[0062] Furthermore, the pathological image database module can deeply analyze the unique pathological characteristics of different conditions, providing a key basis for the subsequent accurate judgment of the type of condition to which the pathological lesion image belongs, and can make different conditions have more refined weight distinctions in the database, improving the accuracy and efficiency of judgment. At the same time, it can accurately quantify the degree of fit between the pathological lesion image and the original pathological characteristics, making the classification and storage of pathological images more scientific and reasonable. Through the matching folder, the doctor's professional experience and subjective judgment ability can be fully utilized to ensure that the pathological lesion image can ultimately be accurately classified into the appropriate disease folder, laying a solid and reliable foundation for subsequent medical analysis and treatment plan formulation.

[0063] Embodiment three:

[0064] Based on the second embodiment, as shown in the accompanying drawings of the specification Figure 1As shown, after determining that a pathological sample has a lesion, the pathological analysis module will transmit the pathological sample information to the pathological image database module, use the pathological image database module to determine the type of disease condition of the pathological sample information, and simultaneously convert the pathological sample information into pathological lesion image information and record it in the corresponding disease condition folder;

[0065] The similar pathology retrieval module can receive the information processed by the pathology analysis module. When judging the disease type of the pathology sample information, it will simultaneously extract the pathology lesion image information in the disease type, and split the pathology sample information and pathology lesion image information into diagnostic components, generate a sample component table and a lesion component table, and then calculate the similarity between the sample component table and the lesion component table. Suppose the component data in the sample component table is Assume that the component data in the lesion component table is Assume that the number of components in the sample component table and the lesion component table is M, and the similarity value is X 相似度 ;

[0066]

[0067] According to the above formula, the similarity value between the sample component table and the lesion component table can be calculated;

[0068] Combined with the 3D reconstruction technology of medical images, the characteristic information of pathological lesion images is fully extracted from a 3D perspective, which can be compared and matched with pathological sample information in a more three-dimensional and accurate manner, thus improving the accuracy and reliability of similarity calculation;

[0069] When generating a sub-treatment plan, the treatment course segmentation module will establish a unit connection function and set default permissions for the unit connection function. Before generating a sub-treatment plan, the treatment course segmentation module will connect different treatment units through the unit connection function and merge them into a treatment unit again. At the same time, the doctor edits the information in the unit connection function so that it can be adjusted according to the doctor's cognition;

[0070] Combined with the medical Internet of Things, when different treatment units are connected and merged through the unit connection function, the operating status and effect of each treatment unit in the patient's body are monitored in real time. Once an abnormality is found in a treatment unit, such as poor drug absorption, abnormal treatment response, etc., the unit connection function is immediately suspended, and the treatment unit and related parts are re-checked and adjusted in detail to ensure the smooth implementation of the entire sub-treatment plan;

[0071] When the stage medical module is combined into a stage treatment plan, it will use the interval time unit and the past efficacy threshold to analyze the maximum time interval between different sub-stage plans, and grant the patient the editing authority of the interval time unit. When the maximum time interval between different sub-stage plans is analyzed, the interval time unit will be restricted according to the maximum time interval between different sub-stage plans. The past efficacy threshold refers to the threshold at which the current sub-stage plan will not affect the treatment effect of the next sub-stage plan after the treatment is completed;

[0072] According to the changing trend of biomarkers during the intervals of different sub-phase programs, combined with the past efficacy threshold, the appropriate interval time can be determined more accurately to avoid blindly treating at fixed time intervals and ignoring the differences in biomarker changes among individual patients, thereby improving the scientificity and effectiveness of the phase treatment program;

[0073] Specific workflow: Use the pathological image database module to determine the disease type of the pathological sample information and record it in the corresponding disease folder, extract the pathological lesion image information in the disease type, split the pathological sample information and pathological lesion image information into diagnostic components, generate a sample component table and a lesion component table, calculate the similarity value of the sample component table and the lesion component table, establish a unit connection function, and set default permissions for it. Before generating a sub-treatment plan, the treatment segmentation module will connect different treatment units through the unit connection function and re-merge them into a treatment unit. The doctor edits the information in the unit connection function, uses the interval time unit and the past efficacy threshold to analyze the maximum time that can be separated between different sub-stage plans, and grants the patient editing permissions for the interval time unit;

[0074] Furthermore, the pathology analysis module can realize the effective linkage between the pathology sample information and the pathology image database, ensure the accuracy and consistency of the judgment of the disease type, and provide a reliable basis for the disease condition for subsequent treatment. The similar pathology retrieval module can provide strong support for determining the appropriate treatment direction and reference. The treatment course segmentation module can flexibly adjust the combination of treatment units according to actual conditions, so that the sub-treatment plan is more in line with the individual differences of patients. The stage medical module not only ensures the continuity and effectiveness of the treatment process, but also fully considers the actual situation and treatment effect of the patient, making the formulation of the stage treatment plan more scientific and reasonable.

[0075] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A blood pathology image analysis system based on artificial intelligence, comprising a patient information acquisition module, the patient information acquisition module is used to obtain the patient's pathological information, collect pathological samples, obtain the patient's blood data, and analyze the patient's disease information based on the blood data, characterized in that: It also includes a pathology image database module, a sample experiment module, a pathology analysis module, a similar pathology retrieval module, and a treatment course segmentation module; The pathological image database module can collect the previous pathological image information, generate pathological lesion images, establish a pathological image database, record the pathological lesion image information in the pathological image database, analyze the difference data between the pathological lesion image information and the normal pathological image information, generate a pathological difference table, mark the pathological difference table on the corresponding pathological image information, and determine the type of disease to which the pathological lesion image information belongs, and establish a disease folder according to the disease type; The sample experiment module extracts the patient's pathological samples and then conducts experimental tests on the pathological samples; The pathological analysis module analyzes whether the pathological sample has lesions. If the pathological sample has lesions, the pathological sample information is analyzed to determine the type of disease. The pathological sample information is then transmitted to the similar pathological retrieval module to generate abnormal sample information. If the pathological sample has no lesions, the pathological sample is marked with a normal label. A similar pathology retrieval module collects abnormal sample information and extracts pathological lesion image information, compares the abnormal sample information with the pathological lesion image information, retrieves the pathological lesion image information with the highest similarity to the abnormal sample information, and then extracts the treatment data of the pathological lesion image information to generate a layer of treatment course information; The treatment course segmentation module extracts the type of disease and analyzes the medical plan for the type of disease based on the type of disease and the first-level treatment course information, thereby generating second-level treatment course information and analyzing its treatment unit. The second-level treatment plan is then separated according to the treatment course unit to generate sub-treatment plans.

2. The blood pathology image analysis system based on artificial intelligence according to claim 1, characterized in that: It also includes a drug medical analysis module and a stage medical module; The drug medical analysis module collects drugs for treating different types of conditions, analyzes the effects of different drugs in different sub-treatment plans, and generates medical effect data; The stage medical treatment module extracts medical plan data, further integrates sub-treatment plans, generates sub-stage plans, combines different sub-stage plans into stage treatment plans, establishes interval time analysis units and past efficacy thresholds, and uses interval time analysis units and past efficacy thresholds to analyze sub-stage plans.

3. The blood pathology image analysis system based on artificial intelligence according to claim 1, characterized in that: When generating the pathological difference table, the pathological image database module will simultaneously analyze the pathological features of different conditions and generate original pathological features. After analyzing the original pathological features of different conditions, the analyzed original pathological features will be recorded in the condition folder, and a weight adding function will be set in the condition folder. Then, the pathological features of the pathological lesion image information will be extracted to generate the current pathological features. The weight adding function information in the condition folder is set to B1, B2, B3, ..., B W , assuming the current pathological features are B1, B2, B3, ..., B T . when When , it can be directly judged that the pathological lesion image of the pathological feature belongs to the disease type corresponding to the weighted addition function information in the disease folder, and no signal will be issued at this time; when When , it can be directly judged that the pathological lesion image of the pathological feature does not belong to the disease type corresponding to the weighted addition function information in the disease folder, and a feature matching signal will be issued; According to the above method, the type of disease to which the pathological lesion image belongs can be preliminarily determined.

4. The blood pathology image analysis system based on artificial intelligence according to claim 3, characterized in that: After the pathological image database module preliminarily determines the type of disease to which the pathological lesion image belongs, it will receive a signal sent when the pathological lesion image is preliminarily determined. After receiving the feature matching signal, it will match the type of disease to which the pathological lesion image belongs through the pathological features. Suppose the original pathological features are Y1, Y2, Y3, ..., Y Z , let the feature matching degree be T 匹配度 , let the summary formula of subset quantity be The feature matching degree of the pathological lesion image can be calculated according to the above formula, and then the corresponding pathological lesion image is recorded in the disease folder corresponding to the original pathological feature with the highest feature matching degree.

5. The blood pathology image analysis system based on artificial intelligence according to claim 4, characterized in that: When calculating the feature matching degree of the pathological lesion image, there will be a situation where the feature matching degree is the same. At this time, a folder with the same matching degree is established, and then the pathological lesion image information with the same feature matching degree is recorded in the folder with the same matching degree, and the disease condition folder with the same feature matching degree is extracted. When the pathological lesion image information exists in the folder with the same matching degree, a judgment signal will be transmitted to the doctor, the doctor will retrieve the content in the folder with the same matching degree, and then the doctor will judge the disease condition folder to which the pathological lesion image information belongs.

6. The blood pathology image analysis system based on artificial intelligence according to claim 1, characterized in that: After determining that a pathological sample has a lesion, the pathological analysis module will transmit the pathological sample information to the pathological image database module, use the pathological image database module to determine the type of disease in the pathological sample information, and simultaneously convert the pathological sample information into pathological lesion image information and record it in the corresponding disease folder.

7. The blood pathology image analysis system based on artificial intelligence according to claim 1, characterized in that: The similar pathology retrieval module can receive information processed by the pathology analysis module, and when judging the type of the pathological sample information, it will simultaneously extract the pathological lesion image information in the disease type, and split the pathological sample information and the pathological lesion image information into diagnostic components, generate a sample component table and a lesion component table, and then calculate the similarity between the sample component table and the lesion component table. Suppose the component data in the sample component table is Assume that the component data in the lesion component table is Assume that the number of components in the sample component table and the lesion component table is M, and the similarity value is X 相似度 ; According to the above formula, the similarity value between the sample component table and the lesion component table can be calculated.

8. The blood pathology image analysis system based on artificial intelligence according to claim 1, characterized in that: When generating a sub-treatment plan, the treatment course segmentation module will establish a unit connection function and set default permissions for the unit connection function. Before generating a sub-treatment plan, the treatment course segmentation module will connect different treatment units through the unit connection function and merge them into a treatment unit. At the same time, the doctor edits the information in the unit connection function so that it can be adjusted according to the doctor's cognition.

9. The blood pathology image analysis system based on artificial intelligence according to claim 2, characterized in that: When the stage medical module is combined into a stage treatment plan, it will use the interval time unit and the past efficacy threshold to analyze the maximum time that can be between different sub-stage plans, and at the same time grant the patient editing authority for the interval time unit. When the maximum time that can be between different sub-stage plans is analyzed, the interval time unit will be restricted according to the maximum time that can be between different sub-stage plans. The past efficacy threshold refers to the threshold at which the current sub-stage plan will not affect the treatment effect of the next sub-stage plan after the treatment is completed.

Images