CN116071303A - Method and device for identifying chronic diabetic nephropathy and storage medium - Google Patents
Method and device for identifying chronic diabetic nephropathy and storage medium Download PDFInfo
- Publication number
- CN116071303A CN116071303A CN202211643281.3A CN202211643281A CN116071303A CN 116071303 A CN116071303 A CN 116071303A CN 202211643281 A CN202211643281 A CN 202211643281A CN 116071303 A CN116071303 A CN 116071303A
- Authority
- CN
- China
- Prior art keywords
- data
- historical
- dkd
- image data
- deep learning
- 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.)
- Withdrawn
Links
- 208000033679 diabetic kidney disease Diseases 0.000 title claims abstract description 78
- 208000007342 Diabetic Nephropathies Diseases 0.000 title claims abstract description 29
- 238000000034 method Methods 0.000 title claims abstract description 22
- 230000001684 chronic effect Effects 0.000 title claims abstract description 20
- 238000013136 deep learning model Methods 0.000 claims abstract description 37
- 210000001525 retina Anatomy 0.000 claims abstract description 27
- 238000012549 training Methods 0.000 claims abstract description 19
- 238000007781 pre-processing Methods 0.000 claims abstract description 17
- DDRJAANPRJIHGJ-UHFFFAOYSA-N creatinine Chemical compound CN1CC(=O)NC1=N DDRJAANPRJIHGJ-UHFFFAOYSA-N 0.000 claims description 26
- 210000002700 urine Anatomy 0.000 claims description 21
- 238000005286 illumination Methods 0.000 claims description 20
- 102000009027 Albumins Human genes 0.000 claims description 14
- 108010088751 Albumins Proteins 0.000 claims description 14
- 229940109239 creatinine Drugs 0.000 claims description 13
- 230000004256 retinal image Effects 0.000 claims description 12
- 208000020832 chronic kidney disease Diseases 0.000 claims description 9
- 210000004369 blood Anatomy 0.000 claims description 8
- 239000008280 blood Substances 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 8
- 210000002966 serum Anatomy 0.000 claims description 7
- 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 claims description 6
- 102000017011 Glycated Hemoglobin A Human genes 0.000 claims description 6
- 108010014663 Glycated Hemoglobin A Proteins 0.000 claims description 6
- 102000015736 beta 2-Microglobulin Human genes 0.000 claims description 6
- 108010081355 beta 2-Microglobulin Proteins 0.000 claims description 6
- 238000013527 convolutional neural network Methods 0.000 claims description 6
- 239000008103 glucose Substances 0.000 claims description 6
- 102000004169 proteins and genes Human genes 0.000 claims description 6
- 108090000623 proteins and genes Proteins 0.000 claims description 6
- 230000009467 reduction Effects 0.000 claims description 6
- 229920006395 saturated elastomer Polymers 0.000 claims description 6
- 230000007547 defect Effects 0.000 abstract description 2
- 238000012216 screening Methods 0.000 description 7
- 238000013135 deep learning Methods 0.000 description 4
- 201000010099 disease Diseases 0.000 description 4
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 4
- 230000008569 process Effects 0.000 description 3
- 206010012689 Diabetic retinopathy Diseases 0.000 description 2
- 108060006698 EGF receptor Proteins 0.000 description 2
- 230000004913 activation Effects 0.000 description 2
- 206010012601 diabetes mellitus Diseases 0.000 description 2
- 230000005182 global health Effects 0.000 description 2
- 230000024924 glomerular filtration Effects 0.000 description 2
- 238000011862 kidney biopsy Methods 0.000 description 2
- 230000003902 lesion Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 201000004569 Blindness Diseases 0.000 description 1
- 206010048554 Endothelial dysfunction Diseases 0.000 description 1
- 206010061218 Inflammation Diseases 0.000 description 1
- 206010027525 Microalbuminuria Diseases 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 238000009535 clinical urine test Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 230000000875 corresponding effect Effects 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000000502 dialysis Methods 0.000 description 1
- 208000016097 disease of metabolism Diseases 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 238000013399 early diagnosis Methods 0.000 description 1
- 208000028208 end stage renal disease Diseases 0.000 description 1
- 201000000523 end stage renal failure Diseases 0.000 description 1
- 230000008694 endothelial dysfunction Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000004054 inflammatory process Effects 0.000 description 1
- 210000003734 kidney Anatomy 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 208000030159 metabolic disease Diseases 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
- 238000005457 optimization Methods 0.000 description 1
- 230000036542 oxidative stress Effects 0.000 description 1
- 230000008506 pathogenesis Effects 0.000 description 1
- 230000001575 pathological effect Effects 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 238000004393 prognosis Methods 0.000 description 1
- 210000001210 retinal vessel Anatomy 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000000758 substrate Substances 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000001225 therapeutic effect Effects 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Images
Classifications
-
- 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
-
- 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
- A61B3/1241—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 specially adapted for observation of ocular blood flow, e.g. by fluorescein angiography
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/18—Eye characteristics, e.g. of the iris
- G06V40/193—Preprocessing; Feature extraction
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
-
- 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]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- Multimedia (AREA)
- Public Health (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- Ophthalmology & Optometry (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Computing Systems (AREA)
- Radiology & Medical Imaging (AREA)
- Biophysics (AREA)
- Software Systems (AREA)
- Veterinary Medicine (AREA)
- Animal Behavior & Ethology (AREA)
- Surgery (AREA)
- Artificial Intelligence (AREA)
- Heart & Thoracic Surgery (AREA)
- Molecular Biology (AREA)
- Hematology (AREA)
- Quality & Reliability (AREA)
- Vascular Medicine (AREA)
- Human Computer Interaction (AREA)
- Data Mining & Analysis (AREA)
- Pathology (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a method and a device for identifying chronic diabetic nephropathy and a readable medium, which comprise the following steps: s1, acquiring historical retina image data and historical DKD data of a DM patient; s2, preprocessing the historical retina image data; s3 inputting the preprocessed historical retina image data and the preprocessed historical DKD data into a deep learning model for training; s4, inputting retina image data and DKD data of the DM patient to be identified into the trained deep learning model, and realizing diabetic nephropathy identification of the DM patient. By adopting the technical scheme of the invention, the defects of invasiveness, high cost and complexity existing in the existing DKD identification grading method are overcome.
Description
Technical Field
The invention belongs to the technical field of disease identification, and particularly relates to a method and a device for identifying chronic diabetic nephropathy and a storage medium.
Background
Diabetes (diabetes mellitus, DM) is a global chronic metabolic disease with which a number of complications are closely related. Diabetic nephropathy (diabetic kidney disease, DKD) is one of the most serious microvascular complications of diabetics, which reduces the quality of life of the patient, increases the economic burden on the global health care system, and increases the mortality rate of diabetics. The DKD is hidden in onset, and once the disease is found to have high severity, the disease is easy to worsen into end-stage renal failure, and the DKD is one of main causes of death of diabetics. Thus, the first and second substrates are bonded together, early screening, early diagnosis, early treatment of DKD for improving the quality of life of diabetics, reducing the occurrence of adverse events, reducing the economic burden of the global health care system, all have great significance. Current screening for DKD relies on the measurement of estimated glomerular filtration rate (estimated glomerular filtration rate, gfr, calculated from serum creatinine) and urine albumin detection. Some pathological markers of DKD have been identified as having significant recognition value for prognosis of DKD patients, where kidney biopsy is a meaningful examination means, but kidney biopsy is not applicable for routine screening.
Similarly, diabetic retinopathy (diabetic retinopathy, DR) is one of the leading causes of blindness in DM patients, as well as microvascular complications of DM. DR is considered one of the recognition criteria for coexisting DKD, as there are studies that find similar histopathological lesions in glomeruli and retinal blood vessels. The pathogenesis of DR and DKD is similar, such as: inflammation, oxidative stress, endothelial dysfunction and microvascular lesions. Although early recognition of DKD is difficult as described above, early recognition of DR by fundus illumination is easy. At present, no comprehensive, universal, noninvasive and low-cost screening method is available for monitoring DM microvascular complications.
Disclosure of Invention
The invention aims to solve the technical problems of providing a method and a device for identifying chronic diabetic nephropathy and a storage medium, and overcomes the defects of invasiveness, high cost and complexity of the existing DKD identification and classification method.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
a method for identifying chronic diabetic nephropathy, comprising the steps of:
s1, acquiring historical retina image data and historical DKD data of a DM patient;
s2, preprocessing the historical retina image data;
s3, inputting the preprocessed historical retina image data and the preprocessed historical DKD data into a deep learning model for training;
s4, inputting retina image data and DKD data of the DM patient to be identified into the trained deep learning model, and realizing diabetic nephropathy identification of the DM patient.
Preferably, the deep learning model is an end-to-end deep learning model based on a ResNet convolutional neural network.
Preferably, the historical DKD data comprises: fasting blood glucose data, glycosylated hemoglobin data, glycosylated albumin data, serum creatinine data, beta 2-microglobulin data, urine creatinine data, urine protein data, urine albumin data.
Preferably, preprocessing the historical retinal image data includes: noise reduction and margin removal processing, removal of saturated pixels of fundus color illumination 255 intensity values, adjustment of fundus color illumination size to 224 x 224 pixels, and marking fundus color illumination according to DM patient UACR and CKD staging
The invention also provides a chronic diabetic nephropathy identification device, which comprises:
an acquisition device for acquiring historical retinal image data and historical DKD data of a DM patient;
the preprocessing device is used for preprocessing the historical retina image data;
the training module is used for inputting the preprocessed historical retina image data and the preprocessed historical DKD data into the deep learning model for training;
the identification module is used for inputting retina image data and DKD data of a DM patient to be identified into the trained deep learning model, so as to realize the identification of diabetic nephropathy of the DM patient.
Preferably, the deep learning model is an end-to-end deep learning model based on a ResNet convolutional neural network.
Preferably, the historical DKD data comprises: fasting blood glucose data, glycosylated hemoglobin data, glycosylated albumin data, serum creatinine data, beta 2-microglobulin data, urine creatinine data, urine protein data, urine albumin data.
Preferably, the preprocessing means preprocesses the history retinal image data includes: noise reduction and margin removal processing, removal of saturated pixels of fundus color illumination 255 intensity values, adjustment of fundus color illumination size to 224 x 224 pixels, and marking fundus color illumination according to DM patient UACR and CKD staging
The present invention also provides a storage medium storing machine-executable instructions that, when invoked and executed by a processor, cause the processor to implement a method of chronic diabetic nephropathy identification.
The invention adopts the Deep Learning (DL) of DR image data to realize the recognition and treatment of DKD. Thus, to assist in formulating therapeutic strategies that promote kidney protection in DM patients, the incidence of DKD and its associated DM mortality are reduced.
Drawings
FIG. 1 is a flow chart of a method of identifying chronic diabetic nephropathy of the present invention;
fig. 2 is a schematic structural view of the chronic diabetic nephropathy identification apparatus of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. The components of the embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.
Example 1:
as shown in fig. 1, the present invention provides a method for identifying chronic diabetic nephropathy, comprising the steps of:
s1, acquiring historical retina image data and historical DKD data of a DM patient;
s2, preprocessing the historical retina image data;
s3, inputting the preprocessed historical retina image data and the preprocessed historical DKD data into a deep learning model for training;
s4, inputting retina image data and DKD data of the DM patient to be identified into the trained deep learning model, and realizing diabetic nephropathy identification of the DM patient.
As an implementation manner of this embodiment, the historical DKD data includes: fasting blood glucose data, glycosylated hemoglobin data, glycosylated albumin data, serum creatinine data, beta 2-microglobulin data, urine creatinine data, urine protein data, urine albumin data.
As one implementation of this embodiment, the historical retinal image data is a color fundus image.
Preprocessing the historical retinal image data includes the following:
1. noise reduction processing and margin cutting processing
2. Removing saturated pixels of the intensity value of the fundus color photograph 255;
3. the size of the fundus color photograph was adjusted to 224×224 pixels according to the requirements of the res net 18.
4. The fundus color illumination was labeled according to the DM patients UACR and CKD stage.
UACR stage is as follows: stage 1 (normal or low albumin urine, UACR < 30 mg/g); stage 2 (microalbuminuria, UACR:30-300 mg/g) and stage 3 (macroalbuminuria, UACR. Gtoreq.300 mg/g). CKD is staged as follows: stage 1 (disease assessment, eGFR is not less than 45mL/min/1.73 m) 2 ) Phase 2 (drug treatment, gfr:15-45mL/min/1.73m 2 ) And phase 3 (dialysis treatment, eGFR)<15mL/min/1.73m 2 )。
As one implementation of this embodiment, training the deep learning model is: according to the historical retina image data and the historical DKD data, obtaining multi-mode depth characteristics through convolution operation; according to the multi-mode depth features, extracting depth features related to the task through convolution and pooling alternately; and obtaining the probability of being identified as different morbidity stages through a weighted loss layer according to the depth characteristics related to the tasks.
The deep learning model is as follows: based on the ResNet convolutional neural network, the end-to-end deep learning model takes historical retinal image data and historical DKD data as inputs and takes six tasks (whether UACR1 period, UACR2 period, UACR3 period, CKD1 period, CKD2 period and CKD3 period) as outputs. The training process of the deep learning model specifically comprises the following steps:
s31, training a ResNet18 model by adopting historical retina image data and historical DKD data, and fine-tuning a network structure and parameters;
s32, using a deep learning random gradient descent (SGD) algorithm as an optimization algorithm of a segmentation model, wherein the learning rate is 0.01, the momentum is 0.9, and each epoch carries out learning rate attenuation with a primary attenuation factor of 1 e-6. In the training process of the deep learning model, the predicted value output by the deep learning model is compared with the real label. When the output predicted value does not coincide with the real tag, the sample is used as a counter-propagating error signal, allowing the network to iteratively adjust its neuron weights to reduce the error.
S33, the inverse of the proportion of the sample is used as a factor in each task to adjust the weight of the neuron, so that the problem of unbalance of the sample is solved, and the accuracy of the deep learning model is improved.
S34, in order to determine the optimal super-parameters of the final deep learning model, a ten-fold cross validation method is used for validating the training data set (the historical retinal image data and the historical DKD data). The scheme randomly divides a training data set into ten independent parts, nine parts are used for training the deep learning model in each operation of the deep learning model, and the rest part is used for testing the deep learning model so as to promote parameter selection and adjustment; ten times are repeated until each part participates in training of the deep learning model.
S35, adopting a weighted loss model, and effectively solving the problem of unbalanced positive and negative proportions of the training sample.
The ResNet18 model is a binary classification model output by the SoftMax classifier, 2 nodes are used in the last layer to generate predicted values, and corresponding UACR and CKD stages are predicted based on fundus illumination.
As an embodiment in this implementation, diabetic nephropathy identification in DM patients can be expressed by high-dimensional feature visualization, and the present example uses Gradient weighted class activation mapping (Gradient-weighted Class Activation Mapping, grad-CAM) to generate a heat map. The target conceptual gradient is used to flow into the final convolution layer, generating a rough localization map to highlight key regions in fundus illumination that are highly correlated with DKD stage. The gradient value represents the magnitude of the contribution of each point in the heat map to the output probability value, and the region with a larger gradient represents a greater influence on the output probability value. After fusion with the original image, the highlighted areas in the heat map represent the areas that are most important to the decision process of the model.
The deep learning model can realize automatic recognition and classification of DKD based on fundus illumination, and can greatly help DM patients. On the one hand, the system can greatly lighten the medical burden: the invention can be added into the existing DR screening system to screen DR and DKD at the same time, thus realizing one-time screening of two DM microvascular complications. Furthermore, for most DM patients, especially in remote areas, periodic screening may be achieved; with the rapid development of artificial intelligence, fundus photography can be realized through a mobile phone in the future to screen DR and DKD. On the other hand, DL mode can reduce economic cost: for medical institutions, reducing medical instrument procurement costs, blood and urine sample processing costs; simple and low cost fundus photography replaces cumbersome and costly multiple blood draws and urine tests for patients.
Example 2:
as shown in fig. 2, the present invention also provides a chronic diabetic nephropathy identification apparatus, comprising:
an acquisition device for acquiring historical retinal image data and historical DKD data of a DM patient;
the preprocessing device is used for preprocessing the historical retina image data;
the training module is used for inputting the preprocessed historical retina image data and the preprocessed historical DKD data into the deep learning model for training;
the identification module is used for inputting retina image data and DKD data of a DM patient to be identified into the trained deep learning model, so as to realize the identification of diabetic nephropathy of the DM patient.
As one implementation of this embodiment, the deep learning model is an end-to-end deep learning model based on a res net convolutional neural network.
As an implementation manner of this embodiment, the historical DKD data includes: fasting blood glucose data, glycosylated hemoglobin data, glycosylated albumin data, serum creatinine data, beta 2-microglobulin data, urine creatinine data, urine protein data, urine albumin data.
As one implementation of this embodiment, the preprocessing device performs preprocessing on the historical retinal image data, including: noise reduction and margin removal processing, removal of saturated pixels of the fundus color illumination 255 intensity values, resizing of the fundus color illumination to 224 x 224 pixels, and marking of the fundus color illumination according to the DM patient UACR and CKD stage.
Example 3:
the present invention also provides a storage medium storing machine-executable instructions that, when invoked and executed by a processor, cause the processor to implement a method of chronic diabetic nephropathy identification.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention.
Claims (9)
1. A method for identifying chronic diabetic nephropathy, comprising the steps of:
s1, acquiring historical retina image data and historical DKD data of a DM patient;
s2, preprocessing the historical retina image data;
s3, inputting the preprocessed historical retina image data and the preprocessed historical DKD data into a deep learning model for training;
s4, inputting retina image data and DKD data of the DM patient to be identified into the trained deep learning model, and realizing diabetic nephropathy identification of the DM patient.
2. The method of claim 1, wherein the deep learning model is an end-to-end deep learning model based on a res net convolutional neural network.
3. The method of claim 1, wherein the historical DKD data comprises: fasting blood glucose data, glycosylated hemoglobin data, glycosylated albumin data, serum creatinine data, beta 2-microglobulin data, urine creatinine data, urine protein data, urine albumin data.
4. The method of claim 1, wherein preprocessing the historical retinal image data comprises: noise reduction and margin removal processing, removal of saturated pixels of the fundus color illumination 255 intensity values, resizing of the fundus color illumination to 224 x 224 pixels, and marking of the fundus color illumination according to the DM patient UACR and CKD stage.
5. A chronic diabetic nephropathy identification apparatus, comprising:
an acquisition device for acquiring historical retinal image data and historical DKD data of a DM patient;
the preprocessing device is used for preprocessing the historical retina image data;
the training module is used for inputting the preprocessed historical retina image data and the preprocessed historical DKD data into the deep learning model for training;
the identification module is used for inputting retina image data and DKD data of a DM patient to be identified into the trained deep learning model, so as to realize the identification of diabetic nephropathy of the DM patient.
6. The chronic diabetic nephropathy identification apparatus of claim 5, wherein the deep learning model is an end-to-end deep learning model based on a res net convolutional neural network.
7. The chronic diabetic nephropathy identification apparatus of claim 6, wherein the historical DKD data comprises: fasting blood glucose data, glycosylated hemoglobin data, glycosylated albumin data, serum creatinine data, beta 2-microglobulin data, urine creatinine data, urine protein data, urine albumin data.
8. The chronic diabetic nephropathy identification apparatus of claim 7, wherein the preprocessing means preprocesses the historical retinal image data comprises: noise reduction and margin removal processing, removal of saturated pixels of the fundus color illumination 255 intensity values, resizing of the fundus color illumination to 224 x 224 pixels, and marking of the fundus color illumination according to the DM patient UACR and CKD stage.
9. A storage medium storing machine executable instructions that when invoked and executed by a processor cause the processor to implement the method of chronic diabetic nephropathy identification of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211643281.3A CN116071303A (en) | 2022-12-20 | 2022-12-20 | Method and device for identifying chronic diabetic nephropathy and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211643281.3A CN116071303A (en) | 2022-12-20 | 2022-12-20 | Method and device for identifying chronic diabetic nephropathy and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116071303A true CN116071303A (en) | 2023-05-05 |
Family
ID=86169229
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211643281.3A Withdrawn CN116071303A (en) | 2022-12-20 | 2022-12-20 | Method and device for identifying chronic diabetic nephropathy and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116071303A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116913524A (en) * | 2023-09-08 | 2023-10-20 | 中国人民解放军总医院第一医学中心 | Method and system for predicting diabetic nephropathy based on retinal vascular imaging |
-
2022
- 2022-12-20 CN CN202211643281.3A patent/CN116071303A/en not_active Withdrawn
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116913524A (en) * | 2023-09-08 | 2023-10-20 | 中国人民解放军总医院第一医学中心 | Method and system for predicting diabetic nephropathy based on retinal vascular imaging |
| CN116913524B (en) * | 2023-09-08 | 2023-12-26 | 中国人民解放军总医院第一医学中心 | Method and system for predicting diabetic nephropathy based on retinal vascular imaging |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111712186B (en) | Method and device for aiding in the diagnosis of cardiovascular disease | |
| US11790645B2 (en) | Diagnosis assistance system and control method thereof | |
| CN108021916B (en) | Deep learning diabetic retinopathy sorting technique based on attention mechanism | |
| CN109635862B (en) | Sorting method for retinopathy of prematurity plus lesion | |
| KR20200005411A (en) | Cardiovascular disease diagnosis assistant method and apparatus | |
| EP3948773A1 (en) | Image-based detection of ophthalmic and systemic diseases | |
| CN111481166A (en) | Automatic identification system based on eye ground screening | |
| Raj et al. | CNN based diabetic retinopathy status prediction using fundus images | |
| CN118675728A (en) | Method and device for aiding in the diagnosis of cardiovascular disease | |
| CN112869697A (en) | Judgment method for simultaneously identifying stage and pathological change characteristics of diabetic retinopathy | |
| CN109464120A (en) | A kind of screening for diabetic retinopathy method, apparatus and storage medium | |
| KR20200139606A (en) | Cervical cancer diagnosis system | |
| CN116071303A (en) | Method and device for identifying chronic diabetic nephropathy and storage medium | |
| CN112957005A (en) | Automatic identification and laser photocoagulation region recommendation algorithm for fundus contrast image non-perfusion region | |
| Gulati et al. | Classification of Diabetic Retinopathy using pre-trained Deep Learning Model-DenseNet 121 | |
| CN111402246A (en) | Eye ground image classification method based on combined network | |
| Heo et al. | Development of a robust eye exam diagnosis platform with a deep learning model | |
| WO2023077238A1 (en) | Hybrid classifier training for feature annotation | |
| CN113989191A (en) | Label migration and classification method for fundus image and contrast image based on deep learning | |
| Kumari et al. | Automated Diabetic Retinopathy Screening With Montage Fundus Images | |
| Ghebrechristos et al. | RetiNet—feature extractor for learning patterns of diabetic retinopathy and age-related macular degeneration from publicly available datasets | |
| Hakeem et al. | Inception V3 and CNN Approach to Classify Diabetic Retinopathy Disease | |
| Abualigah et al. | Hybrid Classification Approach Utilizing DenseUNet+ for Diabetic Macular Edema Disorder Detection. | |
| Kamothi et al. | Automatic Detection of Diabetic Retinopathy Using Convolution Neural Network | |
| Ferdous et al. | An Approach to Classify Ocular Toxoplasmosis Images using Deep Learning Models |
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 | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20230505 |