CN111785363A - AI-guidance-based chronic disease auxiliary diagnosis system - Google Patents
AI-guidance-based chronic disease auxiliary diagnosis system Download PDFInfo
- Publication number
- CN111785363A CN111785363A CN202010494579.7A CN202010494579A CN111785363A CN 111785363 A CN111785363 A CN 111785363A CN 202010494579 A CN202010494579 A CN 202010494579A CN 111785363 A CN111785363 A CN 111785363A
- Authority
- CN
- China
- Prior art keywords
- diagnosis
- guidance
- data
- module
- image
- 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
- 238000003745 diagnosis Methods 0.000 title claims abstract description 86
- 208000017667 Chronic Disease Diseases 0.000 title claims abstract description 36
- 239000000090 biomarker Substances 0.000 claims abstract description 77
- 230000002159 abnormal effect Effects 0.000 claims abstract description 52
- 238000007781 pre-processing Methods 0.000 claims abstract description 39
- 238000012545 processing Methods 0.000 claims abstract description 37
- 238000000034 method Methods 0.000 claims abstract description 32
- 230000003902 lesion Effects 0.000 claims description 38
- 230000011218 segmentation Effects 0.000 claims description 26
- 238000004458 analytical method Methods 0.000 claims description 15
- 238000011002 quantification Methods 0.000 claims description 11
- 238000003384 imaging method Methods 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 8
- 238000000605 extraction Methods 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 5
- 239000008280 blood Substances 0.000 claims description 4
- 210000004369 blood Anatomy 0.000 claims description 4
- 230000036772 blood pressure Effects 0.000 claims description 4
- 210000002216 heart Anatomy 0.000 claims description 4
- 239000011159 matrix material Substances 0.000 claims description 4
- 238000003702 image correction Methods 0.000 claims description 2
- FCKYPQBAHLOOJQ-UHFFFAOYSA-N Cyclohexane-1,2-diaminetetraacetic acid Chemical compound OC(=O)CN(CC(O)=O)C1CCCCC1N(CC(O)=O)CC(O)=O FCKYPQBAHLOOJQ-UHFFFAOYSA-N 0.000 claims 1
- 238000013473 artificial intelligence Methods 0.000 abstract description 30
- 201000010099 disease Diseases 0.000 abstract description 15
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 abstract description 15
- 238000011161 development Methods 0.000 abstract description 5
- 238000004148 unit process Methods 0.000 abstract description 2
- 230000018109 developmental process Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 208000024827 Alzheimer disease Diseases 0.000 description 3
- 206010020772 Hypertension Diseases 0.000 description 3
- 206010039966 Senile dementia Diseases 0.000 description 3
- 206010012601 diabetes mellitus Diseases 0.000 description 3
- 230000004927 fusion Effects 0.000 description 3
- 210000000056 organ Anatomy 0.000 description 3
- 238000012216 screening Methods 0.000 description 3
- 208000024172 Cardiovascular disease Diseases 0.000 description 2
- 210000004204 blood vessel Anatomy 0.000 description 2
- 208000020832 chronic kidney disease Diseases 0.000 description 2
- 238000013145 classification model Methods 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 210000001508 eye Anatomy 0.000 description 2
- 208000030533 eye disease Diseases 0.000 description 2
- 238000010191 image analysis Methods 0.000 description 2
- WQZGKKKJIJFFOK-GASJEMHNSA-N Glucose Natural products OC[C@H]1OC(O)[C@H](O)[C@@H](O)[C@@H]1O WQZGKKKJIJFFOK-GASJEMHNSA-N 0.000 description 1
- 208000017442 Retinal disease Diseases 0.000 description 1
- 206010038923 Retinopathy Diseases 0.000 description 1
- 238000007792 addition Methods 0.000 description 1
- 210000001367 artery Anatomy 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000017531 blood circulation Effects 0.000 description 1
- 210000005252 bulbus oculi Anatomy 0.000 description 1
- 210000001736 capillary Anatomy 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000007499 fusion processing Methods 0.000 description 1
- 239000008103 glucose Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 230000001575 pathological effect Effects 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 210000001525 retina Anatomy 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 210000003462 vein Anatomy 0.000 description 1
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
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B3/00—Apparatus for testing the eyes; Instruments for examining the eyes
- A61B3/10—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions
- A61B3/102—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions for optical coherence tomography [OCT]
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B3/00—Apparatus for testing the eyes; Instruments for examining the eyes
- A61B3/10—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions
- A61B3/12—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions for looking at the eye fundus, e.g. ophthalmoscopes
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B3/00—Apparatus for testing the eyes; Instruments for examining the eyes
- A61B3/10—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions
- A61B3/12—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions for looking at the eye fundus, e.g. ophthalmoscopes
- A61B3/1225—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions for looking at the eye fundus, e.g. ophthalmoscopes using coherent radiation
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B3/00—Apparatus for testing the eyes; Instruments for examining the eyes
- A61B3/10—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions
- A61B3/14—Arrangements specially adapted for eye photography
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4038—Image mosaicing, e.g. composing plane images from plane sub-images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/73—Deblurring; Sharpening
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
-
- 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
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10101—Optical tomography; Optical coherence tomography [OCT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30041—Eye; Retina; Ophthalmic
Landscapes
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Animal Behavior & Ethology (AREA)
- Ophthalmology & Optometry (AREA)
- Heart & Thoracic Surgery (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Veterinary Medicine (AREA)
- Radiology & Medical Imaging (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Primary Health Care (AREA)
- Epidemiology (AREA)
- Quality & Reliability (AREA)
- Pathology (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Eye Examination Apparatus (AREA)
Abstract
The invention provides an AI (artificial intelligence) guidance-based chronic disease auxiliary diagnosis system, and belongs to the technical field of disease auxiliary diagnosis. The AI-guidance-based chronic disease auxiliary diagnosis system comprises a first acquisition unit, a preprocessing unit, an AI algorithm processing unit and an output unit. The first acquisition unit acquires an ophthalmologic image, the preprocessing unit processes the ophthalmologic image into data to be analyzed, the subsequent AI algorithm processing unit is convenient to process, the AI algorithm processing unit analyzes and processes the data to be analyzed into diagnosis information through an AI algorithm, the output unit outputs the diagnosis information, the diagnosis information comprises a diagnosis result and abnormal biological marker parameters, so that a doctor can know the disease type of a patient through the diagnosis result, and meanwhile, the abnormal biological marker parameters can be used for helping the doctor to evaluate the follow-up change trend of a user and predict the development of subsequent focuses, and more comprehensive diagnosis suggestions are provided for the doctor.
Description
Technical Field
The invention belongs to the technical field of auxiliary diagnosis of diseases, and relates to an AI-guidance-based auxiliary diagnosis system for chronic diseases.
Background
The chronic diseases mainly comprise diabetes, hypertension, senile dementia, cardiovascular diseases, chronic kidney diseases and the like, and the death caused by the chronic diseases accounts for about 86 percent of the total death. Chronic diseases are not only harmful to human health, but also have great harm to society and economy, and a great deal of sanitary cost is spent in the diagnosis and treatment of the chronic diseases in every year of China. The fundus retina is the only part of the whole body which can directly and intensively observe arteries, veins and capillaries in a living body, the blood vessels can reflect the dynamic and health condition of the whole body blood circulation of a human, and retinopathy is the characteristic of multiple chronic diseases, so the fundus examination is not only an important method for examining human eye diseases, but also can be used as a window for monitoring multiple diseases such as various chronic diseases and the like, and becomes the standard for early screening and accurate diagnosis. Clinically, doctors in departments such as endocrinology department and neurology department begin to improve the accuracy of chronic diseases diagnosis by means of ophthalmic examination. However, the ophthalmic examination for chronic diseases is just started, and there are problems that the management of patient information, examination and follow-up across departments is still difficult, an ophthalmologist needs to assist a physician in interpreting the fundus examination result of a patient, extra workload is added, and the like. With the rapid development of Artificial Intelligence (AI) in ophthalmology, more and more people are dedicated to research and combine AI technology with ophthalmic disease diagnosis, and the accuracy of AI technology diagnosis is greatly improved and even exceeds that of human doctors. However, existing ophthalmic AI software products focus on diagnosis of a single ophthalmic disease and typically use only a single ophthalmic imaging modality to information fundus color photography or OCT and have not been applied in chronic disease assisted diagnosis. Aiming at the problem that chronic diseases cannot be accurately diagnosed, the close relation between the chronic diseases and eye diseases is considered, and meanwhile, the intelligence of the AI technology is combined, so that the intelligent chronic disease auxiliary diagnosis system based on the guidance of the AI technology is obviously necessary.
Chinese patent CN 107157511a discloses a method and system for diagnosing or screening various human diseases based on image fusion processing of ophthalmology and various organs of human body. The invention can be used for comprehensively storing and managing the comprehensive images and data of the user on a terminal, a local area network and a wide area network. It uses the intelligent analysis of the fusion of the ophthalmology image and each medical image to comprehensively diagnose or screen the human diseases. The invention forms a comprehensive and complete disease screening report of each organ by integrating ophthalmic image analysis, medical image analysis of each organ of a human body and other personal data and utilizing a fusion image processing, machine learning and probability statistical method (called an iNTEGRATE method). The above patent only gives what focus a patient suffers from, and cannot give the size of the focus through the biomarker parameters, which is inconvenient for doctors to understand the diagnosis result.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides an AI-guidance-based chronic disease auxiliary diagnosis system, and the invention aims to solve the technical problems that: how to provide an AI-guidance-based chronic disease auxiliary diagnosis system.
The purpose of the invention can be realized by the following technical scheme:
an AI-guidance-based chronic disease auxiliary diagnosis system comprises a first acquisition unit used for acquiring an ophthalmologic image, a preprocessing unit used for processing the ophthalmologic image into data to be analyzed, an AI algorithm processing unit used for analyzing and processing the data to be analyzed into diagnosis information through an AI algorithm, and an output unit used for outputting the diagnosis information, wherein the diagnosis information comprises a diagnosis result and abnormal biomarker parameters.
Preferably, the first acquisition unit is a fundus camera or a confocal microscope or an OCT imaging apparatus or an oca imaging apparatus.
Preferably, the preprocessing unit comprises a uniform format module for deriving the ophthalmic image acquired by the first acquisition unit and performing uniform format processing on the ophthalmic image into primary data, and a database for storing data.
Preferably, the preprocessing unit further includes a preprocessing module for preprocessing the primary data into data to be analyzed, and the preprocessing includes image rectification and enhancement, image denoising, image motion artifact removal, image registration and stitching.
Preferably, the preprocessing unit performs image rectification and enhancement on the primary data by a Retinex method, the preprocessing unit performs image denoising on the primary data by a low-rank matrix completion method, the preprocessing unit performs image motion artifact removal on the primary data by a two-dimensional transform domain Fourier filtering method, and the preprocessing unit performs image registration and splicing on the primary data by a low-dimensional step mode analysis method.
Preferably, the diagnosis result includes a lesion type and a lesion size, and the AI algorithm processing unit includes a lesion segmentation module for performing a lesion segmentation on the data to be analyzed into a plurality of segmented data, a parameter extraction module for extracting an abnormal biomarker parameter from the segmented data, a diagnosis classification model for performing an analysis process on the abnormal biomarker parameter to obtain a lesion type, and a segmentation quantification model for performing an analysis process on the abnormal biomarker parameter to obtain a lesion size.
Preferably, the diagnosis hierarchical model comprises a first storage module for storing a plurality of types of focus pictures, and a first comparison module for calling the focus pictures in the first storage module and determining the probability that the abnormal biomarker parameter is the target focus type when the similarity between the abnormal biomarker parameter and the preset focus picture is within a type threshold.
Preferably, the segmentation quantification model includes a first statistics module for counting the number of target biomarker parameters with a first threshold of similarity between the abnormal biomarker parameters and a preset lesion picture, a second storage module for storing the lesion size with the number threshold of the preset biomarker parameters matching the corresponding preset lesion type, and a second comparison module for determining the lesion size of the target lesion type corresponding to the target biomarker parameters when the number of the target biomarker parameters is within the number threshold of the preset biomarker parameters in the second storage module.
Preferably, the system further comprises a second acquisition unit for acquiring non-ophthalmic information matched with the ophthalmic image, wherein the non-ophthalmic information comprises blood pressure, heart rate and blood sugar, the first storage module stores the non-ophthalmic information, and the first comparison module determines the probability that the abnormal biomarker parameter is of the target lesion type when the similarity between the abnormal biomarker parameter and the preset lesion picture is within a type threshold and the similarity between the non-ophthalmic information matched with the ophthalmic image and the preset ophthalmic information matched with the preset lesion is within a non-ophthalmic threshold.
Preferably, the AI algorithm processing unit further comprises a training module for training the diagnostic grading model and the segmentation quantification model respectively through the collected data.
The invention has the following beneficial effects: 1. the first acquisition unit acquires an ophthalmologic image, the preprocessing unit processes the ophthalmologic image into data to be analyzed, the subsequent AI algorithm processing unit is convenient to process, the AI algorithm processing unit analyzes and processes the data to be analyzed into diagnosis information through an AI algorithm, the output unit outputs the diagnosis information, and the diagnosis information comprises a diagnosis result and abnormal biomarker parameters, so that a doctor can know the disease type of a patient through the diagnosis result, and meanwhile, the abnormal biomarker parameters can be used for helping the doctor to evaluate the follow-up visit change trend of a user and predict the development of subsequent focuses, and more comprehensive diagnosis suggestions are provided for the doctor;
2. because the formats of the ophthalmic images acquired by various imaging devices are inconsistent and are inconvenient to be processed by a subsequent AI algorithm processing unit, the unified format module derives the ophthalmic images acquired by the first acquisition unit and carries out unified format processing on the ophthalmic images to obtain primary data, so that the diagnosis efficiency is improved, and the database stores the primary data in the unified format and is convenient for subsequent calling;
3. the method comprises the following steps that a preprocessing unit carries out image irradiation correction and enhancement on primary data through a Retinex method, the preprocessing unit carries out image denoising on the primary data through a low-rank matrix completion method, the preprocessing unit carries out image motion artifact removal on the primary data through a two-dimensional transform domain Fourier filtering method, and the preprocessing unit carries out image registration and splicing on the primary data through a low-dimensional step mode analysis method, so that the diagnosis precision is improved;
4. the marking segmentation module is used for marking and segmenting data to be analyzed into a plurality of segmented data, the parameter extraction module is used for extracting abnormal biological marking parameters from the segmented data, the diagnosis grading model is used for analyzing and processing the abnormal biological marking parameters to obtain the type of a focus, and the segmentation quantification model is used for analyzing and processing the abnormal biological marking parameters to obtain the size of the focus, so that a doctor can know what a patient suffers from the type of the focus and the severity of the disease suffered by the patient from the size of the focus, and the diagnosis precision is high;
5. the first storage module stores common ophthalmic pictures of multiple types of focuses, the first comparison module calls the ophthalmic pictures of the focuses and compares the abnormal biomarker parameters with the preset focus pictures one by one, and if the similarity between the abnormal biomarker parameters and the preset focus pictures is within a type threshold value, the probability that the abnormal biomarker parameters are the target focus type can be determined, so that the focus type can be conveniently diagnosed;
6. the second storage module stores a quantity threshold of preset biomarker parameters to be matched with the size of a focus of a corresponding preset focus type, the first statistical module counts the quantity of target biomarker parameters with the similarity between the abnormal biomarker parameters and a preset focus picture as a first threshold, the second comparison module calls the quantity threshold of the preset biomarker parameters stored in the second storage module to be matched with the size of the focus of the corresponding preset focus type after acquiring the quantity of the target biomarker parameters, and determines the focus size of the target focus type corresponding to the target biomarker parameters when the quantity of the target biomarker parameters is within the quantity threshold of the preset biomarker parameters in the second storage module, so that the size of the focus can be conveniently diagnosed;
7. the first storage module stores non-ophthalmological information, and the first comparison module determines the probability that the abnormal biological marking parameter is the target focus type when the similarity between the abnormal biological marking parameter and a preset focus picture is within a type threshold value and the similarity between the non-ophthalmological information matched with an ophthalmological image and the preset ophthalmological information matched with a preset focus is within the non-ophthalmological threshold value, so that the focus type corresponding to the ophthalmological image can be judged through the non-ophthalmological information when the ophthalmological images corresponding to the two focus types are similar, and the diagnosis precision is improved;
8. the training module can respectively train the diagnosis hierarchical model and the segmentation quantitative model through the acquired data, so that the diagnosis precision of the diagnosis hierarchical model and the segmentation quantitative model is improved.
Drawings
FIG. 1 is a schematic view of the structure of the present invention; .
In the figure: 1-a first acquisition unit, 2-a preprocessing unit, 3-an algorithm processing unit, 31-a label segmentation module, 32-a parameter extraction module, 33-a diagnosis hierarchical model, 34-a segmentation quantification model, 35-a training module, 4-an output unit and 5-a second acquisition unit.
Detailed Description
The following are specific embodiments of the present invention and are further described with reference to the drawings, but the present invention is not limited to these embodiments.
Referring to fig. 1, the AI-guided chronic disease auxiliary diagnosis system in the present embodiment includes a first acquisition unit 1 for acquiring an ophthalmic image, a preprocessing unit 2 for processing the ophthalmic image into data to be analyzed, an AI algorithm processing unit 3 for analyzing and processing the data to be analyzed into diagnosis information through an AI algorithm, and an output unit 4 for outputting the diagnosis information, wherein the diagnosis information includes a diagnosis result and an abnormal biomarker parameter.
Here, the first acquisition unit 1 acquires an ophthalmic image, the preprocessing unit 2 processes the ophthalmic image into data to be analyzed, which facilitates the subsequent processing of the AI algorithm processing unit 3, the AI algorithm processing unit 3 analyzes and processes the data to be analyzed into diagnosis information through an AI algorithm, the output unit 4 outputs the diagnosis information, which includes a diagnosis result and an abnormal biomarker parameter, so that a doctor can know the type of a patient suffering from a disease through the diagnosis result, and the abnormal biomarker parameter can be used to help the doctor to evaluate the follow-up visit change trend of the user and predict the development of a subsequent focus, so as to provide a more comprehensive diagnosis suggestion for the doctor, accurately identify a disease, obtain more comprehensive and earlier chronic disease information, treat and treat chronic diseases as early as possible, and help the patient to recover as soon as possible.
The first acquisition unit 1 may be a fundus camera or a confocal microscope or an OCT imaging apparatus or an oca imaging apparatus. Patients with chronic diseases include, but are not limited to, diabetes, hypertension, senile dementia, cardiovascular diseases, and chronic kidney diseases. The first acquisition unit 1 includes, but is not limited to, the above-described devices.
The preprocessing unit 2 may include a unified format module to derive the ophthalmic image acquired by the first acquisition unit 1 and to perform unified format processing on the ophthalmic image into primary data, and a database to store the data. Because the formats of the ophthalmic images acquired by various imaging devices are inconsistent and inconvenient to process by the subsequent AI algorithm processing unit 3, the unified format module derives the ophthalmic images acquired by the first acquisition unit 1 and processes the ophthalmic images into primary data in a unified format, so that the diagnosis efficiency is improved, and the database stores the primary data in the unified format and facilitates subsequent calling. The unified format module unifies the export and storage formats of the acquired data, establishes a PACS system integrating acquisition, input, transmission, classification, storage and reading, and processes to generate a database with a unified storage data format.
The preprocessing unit 2 may further include a preprocessing module for preprocessing the primary data into data to be analyzed, where the preprocessing includes image correction and enhancement, image denoising, image motion artifact removal, image registration and stitching, and improves the diagnosis accuracy.
The preprocessing unit 2 can perform image irradiation correction and enhancement on the primary data through a Retinex method, the preprocessing unit 2 performs image denoising on the primary data through a low-rank matrix completion method, the preprocessing unit 2 performs image motion artifact removal on the primary data through a two-dimensional transform domain Fourier filtering method, and the preprocessing unit 2 performs image registration and splicing on the primary data through a low-dimensional step mode analysis method. The Retinex image enhancement method is adopted to carry out targeted enhancement on blood vessels and nerve structures in the image, sharpens edges and eliminates the problem of low contrast. Adjacent A-scans are aligned A1 to A2, and speckle noise is reduced by applying a vertical translation over A1 to minimize the difference between the two A-scans. And removing the motion artifact by a Fourier filtering method based on a two-dimensional transform domain, and eliminating the image artifact formed by eyeball micromotion in the acquisition process. On the basis of a LoSPA (Low-dimensional Step Pattern Analysis) algorithm, registration algorithm operation of image registration and splicing based on Low-dimensional Step Pattern Analysis is performed by combining human eye information.
The diagnosis result may include a lesion type and a lesion size, and the AI algorithm processing unit 3 includes a marking segmentation module 31 for marking and segmenting data to be analyzed into a plurality of segmented data, a parameter extraction module 32 for extracting an abnormal biomarker parameter in the segmented data, a diagnosis classification model 33 for performing an analysis process on the abnormal biomarker parameter to obtain a lesion type, and a segmentation quantization model 34 for performing an analysis process on the abnormal biomarker parameter to obtain a lesion size. The marking segmentation module 31 performs marking segmentation on data to be analyzed into a plurality of segmented data, the parameter extraction module 32 extracts abnormal biomarker parameters from the segmented data, the diagnosis grading model 33 performs analysis processing on the abnormal biomarker parameters to obtain a focus type, and the segmentation quantification model 34 performs analysis processing on the abnormal biomarker parameters to obtain a focus size, so that a doctor can know what a patient suffers from a focus type and the severity of the patient suffers from the focus size, and the diagnosis precision is high. And extracting the biomarker parameters of various modal images on the segmentation result, comparing the difference of normal people and patients on the biomarker parameters, outputting a judgment basis for improving the understanding of AI diagnosis results by doctors, and evaluating the severity and treatment effect of diseases.
The diagnostic grading model 33 may include a first storage module for storing a plurality of types of lesion pictures, and a first comparison module for calling the lesion pictures in the first storage module and determining a probability that an abnormal biomarker parameter is a target lesion type when a similarity between the abnormal biomarker parameter and a preset lesion picture is within a type threshold. The first storage module stores common ophthalmic pictures of multiple types of focuses, the first comparison module calls the ophthalmic pictures of the focuses and compares the abnormal biomarker parameters with the preset focus pictures one by one, and if the similarity between the abnormal biomarker parameters and the preset focus pictures is within a type threshold value, the probability that the abnormal biomarker parameters are the target focus types can be determined, so that the focus types can be conveniently diagnosed.
The segmentation quantification model 34 may include a first statistic module for counting the number of target biomarker parameters with a first threshold of similarity between the abnormal biomarker parameters and a preset lesion picture, a second storage module for storing the lesion size of which the threshold of the number of the preset biomarker parameters matches the corresponding preset lesion type, and a second comparison module for determining the lesion size of the target lesion type corresponding to the target biomarker parameters when the number of the target biomarker parameters is within the threshold of the number of the preset biomarker parameters in the second storage module. The second storage module stores the number threshold of the preset biomarker parameters to be matched with the size of a focus of a corresponding preset focus type, the first statistical module counts the number of target biomarker parameters with the similarity between the abnormal biomarker parameters and a preset focus picture as a first threshold, the second comparison module calls the number threshold of the preset biomarker parameters stored in the second storage module to be matched with the size of the focus of the corresponding preset focus type after acquiring the number of the target biomarker parameters, and when the number of the target biomarker parameters is within the number threshold of the preset biomarker parameters in the second storage module, the size of the focus of the target focus type corresponding to the target biomarker parameters is determined, so that the size of the focus can be conveniently diagnosed.
The AI-guidance-based chronic disease auxiliary diagnosis system in this embodiment may further include a second acquisition unit 5 configured to acquire non-ophthalmic information matching the ophthalmic image, where the non-ophthalmic information includes blood pressure, heart rate, and blood glucose, the first storage module stores the non-ophthalmic information, and the first comparison module determines the probability that the abnormal biomarker parameter is the target lesion type when the similarity between the abnormal biomarker parameter and the preset lesion picture is within the type threshold and the similarity between the non-ophthalmic information matching the ophthalmic image and the preset lesion is within the non-ophthalmic threshold. The first storage module stores non-ophthalmic information, and the first comparison module determines the probability that the abnormal biomarker parameter is the target focus type when the similarity between the abnormal biomarker parameter and a preset focus picture is within a type threshold value and the similarity between the non-ophthalmic information matched with the ophthalmic image and the preset ophthalmic information matched with the preset focus is within the non-ophthalmic threshold value, so that the focus type corresponding to the ophthalmic image can be judged through the non-ophthalmic information when the ophthalmic images corresponding to the two focus types are similar, and the diagnosis precision is improved. Through the acquired multi-modal ophthalmologic image information and non-ophthalmologic information of a chronic disease diagnosis patient, such as blood pressure, heart rate, blood sugar and the like, deep learning is performed on each modal data respectively, and then model fusion is performed, so that the accuracy of chronic disease auxiliary diagnosis is improved. According to the relevance between the multi-mode image information of the eyes and the chronic diseases, more comprehensive and early chronic disease information is identified by developing an AI algorithm for automatic diagnosis and automatic abnormal pathological feature detection, the provided AI biomarker parameter information can facilitate doctors to effectively evaluate the development trend and treatment effect of the diseases, the multi-mode data can be automatically and rapidly acquired, recorded, transmitted, stored and processed by the AI algorithm, auxiliary diagnosis results are provided for the doctors, and the diagnosis and treatment efficiency and accuracy of the chronic diseases such as diabetes, hypertension and senile dementia are greatly improved.
The AI-algorithm processing unit 3 may further include a training module 35 to train the diagnostic grading model 33 and the segmentation quantification model 34, respectively, by the acquired data. The training module 35 can train the diagnosis hierarchical model 33 and the segmentation quantitative model 34 through the acquired data, so as to improve the diagnosis precision of the diagnosis hierarchical model 33 and the segmentation quantitative model 34, and eliminate the difference after training.
The specific embodiments described herein are merely illustrative of the spirit of the invention. Various modifications or additions may be made to the described embodiments or alternatives may be employed by those skilled in the art without departing from the spirit or ambit of the invention as defined in the appended claims.
Claims (10)
1. An AI-guidance-based chronic disease auxiliary diagnosis system, characterized in that: the device comprises a first acquisition unit (1) for acquiring an ophthalmologic image, a preprocessing unit (2) for processing the ophthalmologic image into data to be analyzed, an AI algorithm processing unit (3) for analyzing and processing the data to be analyzed into diagnosis information through an AI algorithm, and an output unit (4) for outputting the diagnosis information, wherein the diagnosis information comprises a diagnosis result and abnormal biomarker parameters.
2. An AI-guidance-based chronic disease auxiliary diagnostic system according to claim 1, characterized in that: the first acquisition unit (1) is a fundus camera or a confocal microscope or OCT imaging equipment or OCTA imaging equipment.
3. An AI-guidance-based chronic disease auxiliary diagnostic system according to claim 1 or 2, characterized in that: the preprocessing unit (2) comprises a uniform format module and a database, wherein the uniform format module is used for deriving the ophthalmic images acquired by the first acquisition unit (1) and processing the ophthalmic images into primary data in a uniform format.
4. An AI-guidance-based aided diagnosis system of claim 3, wherein: the preprocessing unit (2) further comprises a preprocessing module for preprocessing the primary data into data to be analyzed, wherein the preprocessing comprises image correction and enhancement, image denoising, image motion artifact removal, image registration and splicing.
5. An AI-guidance-based aided diagnosis system of claim 4, wherein: the preprocessing unit (2) corrects and enhances the primary data by a Retinex method, the preprocessing unit (2) denoises the primary data by a low-rank matrix completion method, the preprocessing unit (2) removes image motion artifacts from the primary data by a two-dimensional transform domain Fourier filtering method, and the preprocessing unit (2) performs image registration and splicing on the primary data by a low-dimensional step mode analysis method.
6. An AI-guidance-based aided diagnosis system of claim 4, wherein: the diagnosis result comprises a focus type and a focus size, and the AI algorithm processing unit (3) comprises a mark segmentation module (31) for performing mark segmentation on data to be analyzed into a plurality of segmented data, a parameter extraction module (32) for extracting abnormal biomarker parameters from the segmented data, a diagnosis grading model (33) for performing analysis processing on the abnormal biomarker parameters to obtain the focus type, and a segmentation quantification model (34) for performing analysis processing on the abnormal biomarker parameters to obtain the focus size.
7. An AI-guidance-based aided diagnosis system of claim 6, wherein: the diagnosis grading model (33) comprises a first storage module for storing a plurality of types of focus pictures, and a first comparison module for calling the focus pictures in the first storage module and determining the probability that the abnormal biological marking parameter is the target focus type when the similarity between the abnormal biological marking parameter and the preset focus picture is within a type threshold.
8. An AI-guidance-based chronic disease auxiliary diagnostic system according to claim 7, characterized in that: the segmentation quantification model (34) comprises a first statistic module for counting the number of target biomarker parameters with the similarity between the abnormal biomarker parameters and a preset lesion picture as a first threshold, a second storage module for storing the lesion size of which the number threshold of the preset biomarker parameters is matched with the corresponding preset lesion type, and a second comparison module for determining the lesion size of the target lesion type corresponding to the target biomarker parameters when the number of the target biomarker parameters is within the number threshold of the preset biomarker parameters in the second storage module.
9. An AI-guidance-based chronic disease auxiliary diagnostic system according to claim 8, characterized in that: the non-ophthalmological information acquisition system is characterized by further comprising a second acquisition unit (5) used for acquiring non-ophthalmological information matched with an ophthalmological image, wherein the non-ophthalmological information comprises blood pressure, heart rate and blood sugar, the first storage module stores the non-ophthalmological information, and the first comparison module determines the probability that the abnormal biomarker parameter is of the target lesion type when the similarity between the abnormal biomarker parameter and a preset lesion picture is within a type threshold value and the similarity between the non-ophthalmological information matched with the ophthalmological image and the preset ophthalmological information matched with the preset lesion is within a non-ophthalmological threshold value.
10. An AI-guidance-based chronic disease auxiliary diagnostic system according to claim 8, characterized in that: the AI algorithm processing unit (3) further comprises a training module (35) to train the diagnostic grading model (33) and the segmentation quantification model (34) respectively by means of the acquired data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010494579.7A CN111785363A (en) | 2020-06-03 | 2020-06-03 | AI-guidance-based chronic disease auxiliary diagnosis system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010494579.7A CN111785363A (en) | 2020-06-03 | 2020-06-03 | AI-guidance-based chronic disease auxiliary diagnosis system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111785363A true CN111785363A (en) | 2020-10-16 |
Family
ID=72753509
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010494579.7A Pending CN111785363A (en) | 2020-06-03 | 2020-06-03 | AI-guidance-based chronic disease auxiliary diagnosis system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111785363A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112820398A (en) * | 2021-01-22 | 2021-05-18 | 神威超算(北京)科技有限公司 | Artificial intelligence-based lung diagnosis auxiliary system and method |
| CN112908466A (en) * | 2021-01-13 | 2021-06-04 | 中国科学院计算技术研究所 | Alzheimer disease intelligent decision support method and system based on transfer learning |
| CN112992336A (en) * | 2021-02-05 | 2021-06-18 | 山西医科大学 | Intelligent pathological diagnosis system |
| CN115578554A (en) * | 2021-06-21 | 2023-01-06 | 数坤(北京)网络科技股份有限公司 | Blood vessel focus identification method, device, electronic equipment and readable storage medium |
| CN116046643A (en) * | 2023-03-30 | 2023-05-02 | 深圳市帝迈生物技术有限公司 | Auxiliary diagnostic information providing device and blood analysis system |
| WO2023186051A1 (en) * | 2022-03-31 | 2023-10-05 | 深圳市帝迈生物技术有限公司 | Auxiliary diagnosis method and apparatus, and construction apparatus, analysis apparatus and related product |
| CN117389529A (en) * | 2023-12-12 | 2024-01-12 | 神州医疗科技股份有限公司 | AI interface calling method and system based on PACS system |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160305951A1 (en) * | 2015-04-16 | 2016-10-20 | Po-Kang Lin | Biomarker for human eyes and method thereof |
| US20180196924A1 (en) * | 2017-01-09 | 2018-07-12 | International Business Machines Corporation | Computer-implemented method and system for diagnosis of biological conditions of a patient |
| CN108682015A (en) * | 2018-05-28 | 2018-10-19 | 科大讯飞股份有限公司 | Lesion segmentation method, apparatus, equipment and storage medium in a kind of biometric image |
| CN109829882A (en) * | 2018-12-18 | 2019-05-31 | 苏州比格威医疗科技有限公司 | A kind of stages of DR prediction technique |
| CN110010219A (en) * | 2019-03-13 | 2019-07-12 | 杭州电子科技大学 | Optical coherence tomography image retinopathy intelligent checking system and detection method |
-
2020
- 2020-06-03 CN CN202010494579.7A patent/CN111785363A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160305951A1 (en) * | 2015-04-16 | 2016-10-20 | Po-Kang Lin | Biomarker for human eyes and method thereof |
| US20180196924A1 (en) * | 2017-01-09 | 2018-07-12 | International Business Machines Corporation | Computer-implemented method and system for diagnosis of biological conditions of a patient |
| CN108682015A (en) * | 2018-05-28 | 2018-10-19 | 科大讯飞股份有限公司 | Lesion segmentation method, apparatus, equipment and storage medium in a kind of biometric image |
| CN109829882A (en) * | 2018-12-18 | 2019-05-31 | 苏州比格威医疗科技有限公司 | A kind of stages of DR prediction technique |
| CN110010219A (en) * | 2019-03-13 | 2019-07-12 | 杭州电子科技大学 | Optical coherence tomography image retinopathy intelligent checking system and detection method |
Non-Patent Citations (3)
| Title |
|---|
| 中国医药教育协会智能医学专委会智能眼科学组等: "基于眼底照相的糖尿病视网膜病变人工智能筛查系统应用指南", 《中华实验眼科杂志》 * |
| 李梭: "肺部CT影像慢阻肺病灶三维定量化分析研究", 《中国优秀博硕士学位论文全文数据库(博士)信息科技辑》 * |
| 龚雁等: "基于深度学习OCT辅助诊断湿性年龄相关性黄斑变性算法的应用", 《中华实验眼科杂志》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112908466A (en) * | 2021-01-13 | 2021-06-04 | 中国科学院计算技术研究所 | Alzheimer disease intelligent decision support method and system based on transfer learning |
| CN112908466B (en) * | 2021-01-13 | 2023-08-04 | 中国科学院计算技术研究所 | Alzheimer's disease intelligent decision support method and system based on transfer learning |
| CN112820398A (en) * | 2021-01-22 | 2021-05-18 | 神威超算(北京)科技有限公司 | Artificial intelligence-based lung diagnosis auxiliary system and method |
| CN112992336A (en) * | 2021-02-05 | 2021-06-18 | 山西医科大学 | Intelligent pathological diagnosis system |
| CN115578554A (en) * | 2021-06-21 | 2023-01-06 | 数坤(北京)网络科技股份有限公司 | Blood vessel focus identification method, device, electronic equipment and readable storage medium |
| CN115578554B (en) * | 2021-06-21 | 2024-02-02 | 数坤(上海)医疗科技有限公司 | Vascular focus identification method, device, electronic equipment and readable storage medium |
| WO2023186051A1 (en) * | 2022-03-31 | 2023-10-05 | 深圳市帝迈生物技术有限公司 | Auxiliary diagnosis method and apparatus, and construction apparatus, analysis apparatus and related product |
| CN116046643A (en) * | 2023-03-30 | 2023-05-02 | 深圳市帝迈生物技术有限公司 | Auxiliary diagnostic information providing device and blood analysis system |
| CN116046643B (en) * | 2023-03-30 | 2024-01-16 | 深圳市帝迈生物技术有限公司 | Auxiliary diagnostic information providing device and blood analysis system |
| CN117389529A (en) * | 2023-12-12 | 2024-01-12 | 神州医疗科技股份有限公司 | AI interface calling method and system based on PACS system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111785363A (en) | AI-guidance-based chronic disease auxiliary diagnosis system | |
| Cao et al. | Hierarchical method for cataract grading based on retinal images using improved Haar wavelet | |
| Sánchez et al. | Retinal image analysis to detect and quantify lesions associated with diabetic retinopathy | |
| CN111986211A (en) | Deep learning-based ophthalmic ultrasonic automatic screening method and system | |
| CN109948719B (en) | Automatic fundus image quality classification method based on residual dense module network structure | |
| Kauppi | Eye fundus image analysis for automatic detection of diabetic retinopathy | |
| CN113962311A (en) | Knowledge data and artificial intelligence driven ophthalmic multi-disease identification system | |
| WO2020042500A1 (en) | Health big data service method and system based on remote fundus screening | |
| CN112950737B (en) | Fundus fluorescence contrast image generation method based on deep learning | |
| CN113768461B (en) | Fundus image analysis method, fundus image analysis system and electronic equipment | |
| Karthiyayini et al. | Retinal image analysis for ocular disease prediction using rule mining algorithms | |
| Kumar et al. | Computational intelligence in eye disease diagnosis: a comparative study | |
| Giancardo | Automated fundus images analysis techniques to screen retinal diseases in diabetic patients | |
| KR20230144517A (en) | Method for predicting cardiovascular disease using eye image | |
| Preity et al. | Automated computationally intelligent methods for ocular vessel segmentation and disease detection: a review | |
| KR20200011530A (en) | Method for predicting cardio-cerebrovascular disease using eye image | |
| Noronha et al. | A review of fundus image analysis for the automated detection of diabetic retinopathy | |
| Chandra et al. | Confocal Corneal Endothelium Dystrophy’s Analysis using Particle Filter | |
| Kubde et al. | Automated Computer Aided Detection of Diabetic Retinopathy Using Machine Learning Hybrid Model | |
| Rozlan et al. | A proposed diabetic retinopathy classification algorithm with statistical inference of exudates detection | |
| US20240281968A1 (en) | System and method for retinal optical coherence tomography classification using region-of-interest aware resnet | |
| US20220151482A1 (en) | Biometric ocular measurements using deep learning | |
| US20240245291A1 (en) | An intelligent single tool for clinicians | |
| Soomro et al. | A Wide-Ranging Review on Diabetic Retinotherapy | |
| Kaur et al. | 34 Images Using Deep Learning: Methods |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20201016 |