CN111080643A - Method and device for classifying diabetes and related diseases based on fundus images - Google Patents
Method and device for classifying diabetes and related diseases based on fundus images Download PDFInfo
- Publication number
- CN111080643A CN111080643A CN201911421191.8A CN201911421191A CN111080643A CN 111080643 A CN111080643 A CN 111080643A CN 201911421191 A CN201911421191 A CN 201911421191A CN 111080643 A CN111080643 A CN 111080643A
- Authority
- CN
- China
- Prior art keywords
- information
- diabetes
- machine learning
- learning model
- loss value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine 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
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30041—Eye; Retina; Ophthalmic
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- Medical Informatics (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Quality & Reliability (AREA)
- Radiology & Medical Imaging (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Eye Examination Apparatus (AREA)
Abstract
The invention provides a classification method and equipment for diabetes and related diseases based on fundus images, wherein the method comprises the steps of obtaining sample data, wherein the sample data comprises fundus images and a plurality of types of information, and the plurality of types of information comprise diabetes type information and at least one type of diabetes related disease information, or at least two types of diabetes related disease type information; training a machine learning model by using a large amount of the sample data to output an evaluation result, wherein the evaluation result at least comprises a classification result corresponding to one type of the type information, and the machine learning model comprises a feature extraction network and at least one output network, wherein the feature extraction network is used for extracting feature information from the fundus image, and the at least one output network is used for outputting the evaluation result according to the feature information respectively; and the machine learning model at least adjusts the parameters of the machine learning model according to the output evaluation result and the type information in the sample data.
Description
Technical Field
The invention relates to the field of medical image analysis, in particular to a method and equipment for classifying diabetes and related diseases based on fundus images.
Background
In recent years, machine learning techniques have been widely used in the medical field, and in particular, machine learning techniques typified by deep learning have been attracting attention in the medical imaging field. In the aspect of fundus image detection, the deep learning technology can accurately detect a certain characteristic of a fundus image, for example, a deep learning model is trained by using a large number of fundus image samples of diabetics, and the trained model can be used for detecting diabetes of the fundus image.
Chinese patent application No. 201810387302.7 discloses a fundus image detection method based on machine learning, which performs feature detection with high significance on the entire region of a fundus image to thereby complete apparent disease feature recognition, and then performs feature detection with low significance on a specific region to complete further disease feature recognition. The method classifies fundus images with various characteristics greatly, performs fine detection of subareas in images without obvious characteristics, performs detection in series step by step, independently outputs detection results, and can realize accurate detection of obvious characteristics and tiny characteristics at the same time. The scheme is suitable for simultaneously identifying a plurality of disease types, for example, whether the fundus image has a plurality of irrelevant disease characteristics such as macular hole, glaucoma and diabetes is detected to obtain corresponding classification, and the method is not suitable for the relevant disease classification.
Disclosure of Invention
In view of the above, the present invention provides a method for constructing a classification model of diabetes and related diseases, comprising:
acquiring sample data which comprises fundus images and a plurality of types of information, wherein the plurality of types of information are diabetes type information and at least one type of diabetes-related disease information, or at least two types of diabetes-related disease type information;
training a machine learning model by using a large amount of the sample data to output an evaluation result, wherein the evaluation result at least comprises a classification result corresponding to one type of the type information, and the machine learning model comprises a feature extraction network and at least one output network, wherein the feature extraction network is used for extracting feature information from the fundus image, and the at least one output network is used for outputting the evaluation result according to the feature information respectively; and the machine learning model at least adjusts the parameters of the machine learning model according to the output evaluation result and the type information in the sample data.
Optionally, there are a plurality of the output networks, and the output networks respectively output classification results corresponding to different types of information according to the feature information; the machine learning model adjusts parameters of the machine learning model at least according to the output evaluation result and type information in the sample data, and the method comprises the following steps:
determining a second loss value according to the difference between the classification result and the corresponding type information;
determining a first loss value according to the second loss value;
and the machine learning model adjusts self parameters according to the first loss value.
Optionally, the sample data further includes auxiliary information, and the auxiliary information includes diabetes related information and/or diabetes-associated disease related information.
Optionally, there are a plurality of output networks, including an output network for outputting classification results corresponding to different types of information according to the feature information, and an output network for outputting identification results corresponding to the auxiliary information according to the feature information; the evaluation result includes a classification result corresponding to each of the types of information and an identification result corresponding to the auxiliary information;
the machine learning model adjusts parameters of the machine learning model according to the output evaluation result and type information in the sample data, and the method comprises the following steps:
determining a second loss value according to the difference between the classification result and the corresponding type information;
determining a third loss value according to the difference between the auxiliary information and the corresponding recognition result;
determining a first loss value according to the second loss value and the third loss value;
and the machine learning model adjusts self parameters according to the first loss value.
Optionally, the machine learning model adjusts a parameter of the feature extraction network at least according to the first loss value.
Optionally, the machine learning model adjusts a parameter of the corresponding output network according to the second loss value.
Optionally, the machine learning model adjusts a parameter of the corresponding output network according to the second loss value and the third loss value.
Optionally, the diabetes-associated disease comprises a diabetic complication and/or diabetic complication.
The invention also provides a method for classifying diabetes and related diseases, which comprises the following steps: acquiring a fundus image of a user; and identifying the fundus image by using the machine learning model constructed by the method, and outputting at least one classification result.
Accordingly, the present invention provides a diabetes classification model construction apparatus, comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the processor, the instructions being executable by the at least one processor to cause the at least one processor to perform the method of classification model construction for diabetes and related diseases as described above.
Accordingly, the present invention provides an apparatus for classifying diabetes and related diseases, comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to cause the at least one processor to perform the method of classifying diabetes and related conditions described above.
According to the method and the device for classifying the diabetes and the related diseases, provided by the invention, the eyeground image of the user is acquired, the characteristics of the eyeground image are extracted by using the machine learning model, and the classification result for representing the diabetes and the related diseases of the user can be output according to the characteristics. The scheme does not need to collect blood samples or other body indexes of a user, only needs to collect fundus images, realizes a noninvasive classification process, and optimizes classification performance due to the introduction of various types of diabetes and related diseases into the model in the training process, so that the output classification result is more accurate.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a schematic diagram of a machine learning model training in an embodiment of the present invention;
FIG. 2 is a diagram illustrating another machine learning model training in an embodiment of the present invention;
FIG. 3 is a diagram illustrating an embodiment of a machine learning model for performing diabetes and related disease classification;
FIG. 4 is a diagram of another machine learning model for performing diabetes and related disease classification according to an embodiment of the present invention.
Detailed Description
The technical solutions of the present invention will be described clearly and completely with reference to the accompanying drawings, and it should be understood that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In addition, the technical features involved in the different embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
One embodiment of the present invention provides a method for constructing a classification model of diabetes and related diseases, which relates to a machine learning model, and as shown in fig. 1, the model comprises a feature extraction network 11 and a plurality of output networks 12. The network described in the present application refers to a neural network, in particular, a convolutional neural network. Both the classification network and the regression network described below require a feature extraction structure consisting of a convolutional layer, a pooling layer, and an activation layer. One or more feature extraction structures may form a feature extraction network, and the network extracted features may be input into a subsequent output network. The network has initialized parameters that are trained in this embodiment using sample data to optimize the parameters, thereby enabling the network to output classification results for diabetes and related diseases.
First, sample data including a fundus image and a plurality of types of information is acquired. The plurality of types of information comprises two conditions, wherein the first condition comprises one diabetes type information and a plurality of diabetes-related disease type information; the second case is to include a plurality of diabetes-associated disease type information, but not diabetes type information. The diabetes-related diseases mentioned in the present application include both the concepts of complications and complications in the medical domain, such as diabetic hypertension, diabetic retinopathy, cataracts, cardiovascular diseases, diabetic nephropathy, all belonging to the diabetes-related diseases mentioned in the present application. Such as more than 100 of the diabetes-associated diseases mentioned above, the type information may include any portion or all thereof.
By way of example, the diabetes type information may be at least two types of information, i.e., with diabetes or without diabetes, and may further indicate a specific type of diabetes, such as type 1 diabetes, type 2 diabetes; similarly, the diabetes-associated disease type information can be at least the presence, absence, or further indicative of a particular type of disease.
Sample data should be taken from the real subject, such as the above type of information of a person can be obtained by reading physical examination data or case data. The fundus image is an image taken by a fundus camera, and includes images of organs such as the macula lutea, optic disc, and blood vessels. The retinal blood vessels of the fundus are blood vessels visible in the human body and are considered as windows for understanding the blood vessels of other organs. Therefore, the characteristics of the fundus blood vessels reflect the state of some organs to some extent, and fundus images are very closely linked to diabetes and related diseases.
In practical applications, some preprocessing may be performed on the fundus image, for example, edges may be cropped, the size may be adjusted, the contrast may be enhanced, and the like, to normalize the data content and improve the image quality. The fundus image and the plurality of types of information in one sample data should be from the same subject. To train a machine learning model, a large amount of sample data should be acquired and divided into training data and test data.
And then, training the machine learning model by using a large amount of sample data, and enabling the machine learning model to output an evaluation result. In this embodiment, the evaluation result output by the model at least includes classification results corresponding to each type of information, that is, the type information in the sample data is used as a tag, and the output content of the model corresponds to the tag. For example, if one sample data includes a fundus image P and type information a, type information B, and type information C, the fundus image P has A, B, C three labels, and the model recognizes the fundus image P and outputs a classification result a ' corresponding to a, a classification result B ' corresponding to B, and a classification result C ' corresponding to C.
The feature extraction network 11 shown in fig. 1 is for extracting feature information from the fundus image, and the plurality of output networks 12 are for outputting evaluation results corresponding to the type information based on the feature information, respectively. The output networks share the same characteristic extracted by the extraction network, and output the result corresponding to the label. For example, the feature extraction network 11 extracts the fundus image P to obtain feature information featurepAnd the first output network is according to featurepOutputting the classification result A 'corresponding to the label A, and outputting the classification result A' by the second output network according to featurepOutputting score corresponding to label BClass result B' and third output network according to featurepAnd outputting a classification result C' corresponding to the label C.
And the machine learning model adjusts the parameters of the machine learning model according to the difference between the output evaluation result and the type information in the sample data. According to the characteristics of the neural network, the model needs to determine a loss value (loss) according to the difference, and optimizes the parameters of the model by reverse transmission so as to reduce the difference.
According to the method for constructing the classification model of diabetes and related diseases, the model with learning capability consisting of the neural network is trained by acquiring the sample data consisting of the fundus image, the type information of various diabetes and related diseases thereof, the evaluation result output by the model is compared with the label of the sample data in the training process, the parameter of the model is optimized according to the difference, so that the model can learn the relationship between the type information of the diabetes and the related diseases thereof and the content presented by the fundus image, and the classification result of the diabetes and the related diseases thereof can be obtained through the fundus image by the constructed model.
In this embodiment, all the evaluation results belong to classification results, including two or more classifications, that is, all the output networks may be classification networks. The output network (or output structure) of a classification network typically contains an output layer of a Softmax or Sigmoid function, with the output typically being a confidence or probability of 0-1 to describe the probability or confidence that the input belongs to a certain class or classes.
Preferably, the difference between the classification result and the label can be represented using a cross-entropy function, such as
According to the model structure in this embodiment, each classification nodeThe difference between an effect and its corresponding label can be used to calculate a loss value, such as the loss value L for the type of diabetesDiabetes mellitusLoss value L of diabetic hypertensionHypertension (hypertension)And loss value L of diabetic nephropathyRenal diseaseAnd loss value L of diabetic cardiovascular diseaseCardiovascular diseasesAnd so on, these loss values corresponding to the type information may be referred to as second loss values, which are plural. When the model adjusts the parameters according to these loss values, a total loss value (first loss value) L can be calculated according to these loss valuestotI.e. Ltot＝f(LDiabetes mellitus，LHypertension (hypertension)，LRenal disease，LCardiovascular diseases…) and the degree of influence of these second loss values on the total loss value may be different, for example, the influence weights of the respective loss values may be set, and the total loss value may be calculated by linear or non-linear weighting. The final model adjusts the parameters according to the calculated total loss value.
Further, the machine learning model adjusts parameters of the feature extraction network 11 at least according to the first loss value, so that feature extraction performance is optimized, feature information obtained by the feature extraction performance is more accurate, and accuracy of an output network is improved. The parameters of the respective output network 12 may also be adjusted according to the first loss value, or the parameters of the respective output network may be adjusted according to a plurality of second loss values. For example, for an output network of diabetes type, the loss value L may be based on the diabetes typeDiabetes mellitusAdjusting parameters of the network, for an output network of the diabetic nephropathy type, according to the loss value L of the diabetic nephropathyNephropathy BParameters of the network are adjusted.
The above preferred embodiment optimizes the parameters of the feature extraction network by the total loss value, and optimizes the parameters of the corresponding output network by the loss values corresponding to the various evaluation results, thereby providing the efficiency and performance of the model training.
In an alternative embodiment, the sample data further comprises auxiliary information, the auxiliary information comprising diabetes related information and/or diabetes related disease related information, such as age information, weight information, blood pressure information, some related medical history information, and the like. Reference may be made in particular to the following table:
the table shows the correspondence of some optional auxiliary information to the type information. In this embodiment, a model may be trained using any one or more of the above-mentioned auxiliary information, which also serves as a label, and some of the output networks 12 in the model are used to output a recognition result corresponding to the above-mentioned auxiliary information based on the feature information of the fundus image. For example, if one sample data includes a fundus image P and type information a, type information B, auxiliary information C, and auxiliary information D, the fundus image P has A, B, C, D four labels, and the model recognizes the fundus image P and outputs a classification result a 'corresponding to a, a classification result B' corresponding to B, a recognition result C 'corresponding to C, and a recognition result D' corresponding to D.
Such an output network may be a classification network or a regression network, the outcome of which is a numerical value. The output of the regression network is a numerical prediction value for a specific index, such as: age, sex, systolic blood pressure, etc. The output network of the network at least comprises a fully-connected layer for weighting and outputting the input, and an activation layer can be added to change the response value of the input, such as outputting a value less than 0 to 0 and outputting a value greater than 0 to be unchanged by using Relu as an activation function. For example, if an auxiliary information is age, i.e., a tag is an age value, the output network will perform regression prediction based on the characteristic information and output a value, i.e., the predicted age. The evaluation result output by the output network is a result corresponding to the auxiliary information, and the recognition targets can be converted into a classification problem or a regression prediction problem, for example, the age can be a regression prediction result, or the age can be segmented and converted into a classification result.
The difference between the regression prediction result and the label can be expressed by using an Error function, such as Mean square Error function (MSE), Mean Absolute Error function (MAE), Mean Absolute Percentage Error function (MAPE), and the like. Thus, the loss value of each regression prediction result and the corresponding label can be obtained.
The evaluation result of this embodiment may include both the classification result and the regression prediction result, and the plurality of output networks include both the classification network and the regression prediction network, and these two results respectively express the difference between them and the label by the above-mentioned method, thereby determining each loss value. According to the preferred scheme, the proper types of networks are respectively adopted for various types of labels and output, and the accuracy of the training result is improved by combining the advantages of a classification network and a regression network.
With respect to the loss function, according to the model structure in the embodiment, a loss value, such as a loss value L of age, can be calculated for each difference between the regression prediction result and its corresponding auxiliary informationageLoss value L of systolic pressureSBPDiastolic blood pressure loss value LDBPSex loss value LgenderAnd so on, these loss values corresponding to the side information may be referred to as third loss values, and there may be one or more of the third loss values. In combination with the second loss value related to the classification result, when the model adjusts the parameters, a total loss value (first loss value) L can be calculated according to the second loss value and the third loss valuetotI.e. Ltot＝f(Lage，LSBP，LDBP，Lgender，LDiabetes mellitus，LHypertension (hypertension)，LRenal disease，LCardiovascular diseases…), and the influence of these third and second loss values on the total loss value may be different, for example, the influence weight of each loss value may be set, and the total loss value may be calculated by linear or non-linear weighting. The final model adjusts the parameters according to the calculated total loss value.
Further, the machine learning model adjusts parameters of the feature extraction network 11 at least according to the first loss value, so that feature extraction performance is optimized, feature information obtained by the feature extraction performance is more accurate, and accuracy of an output network is improved. The parameters of the respective output network 12 may also be dependent on the firstAnd adjusting the loss value, or adjusting the parameter of the corresponding output network according to a plurality of second loss values or third loss values respectively. For example, for an output network of diabetes type, the loss value L may be based on the diabetes typeDiabetes mellitusAdjusting parameters of the network, for the output network of the systolic pressure, according to the loss value L of the systolic pressureSBPParameters of the network are adjusted.
In another embodiment of the present invention, a machine learning model is provided, as shown in fig. 2, which includes a feature extraction network 11 and a unique output network 12, for outputting a classification result according to feature information extracted by the feature extraction network 11 and a plurality of types of information, and adjusting parameters thereof according to a loss value.
The sample data used to train such a model still includes the fundus image and a plurality of types of information, one of which serves as a label and the other as input data to the output network. For example, one sample data includes a fundus image P having a label of a, type information B, and type information C, and the feature extraction network 11 extracts the fundus image P to obtain feature information featurepAnd output network 12 is according to B, C, featurepAnd outputting a classification result A' corresponding to the label A.
The machine learning model of the embodiment has a unique output network, takes the type information of a plurality of diabetes and related diseases as input, and enables the model to learn the relation between the fundus image and a certain type based on the plurality of types of information, thereby improving the classification efficiency and accuracy.
In a preferred embodiment, the sample data may further include the auxiliary information, which is also used as input data of the output network. For example, one sample data includes a fundus image P having a label of a, type information B, auxiliary information C, and auxiliary information D, and the feature extraction network 11 extracts the fundus image P to obtain feature information featurepAnd output network 12 is according to B, C, D, featurepAnd outputting a classification result A' corresponding to the label A.
The machine learning model of the present embodiment has a unique output network, and uses type information and auxiliary information of a plurality of types of diabetes and related diseases thereof, so that the model learns the relationship between the fundus image and a certain type based on the plurality of types of information and auxiliary information, whereby the classification efficiency and accuracy can be improved.
One embodiment of the present invention provides a diabetes classification method, which relates to a machine learning model, and the model can be obtained by training using the method described in the above embodiment. Referring to fig. 3, the method of the present embodiment includes: a fundus image 20 of a user is acquired, the fundus image 20 is recognized by a machine learning model 21, and a classification result is output. The classification result of the present embodiment includes a plurality of types, such as diabetes type information and at least one type of diabetes-related disease information, or at least two types of diabetes-related disease information.
The machine learning model 21 includes a feature extraction network 211 and a plurality of output networks 212, wherein the feature extraction network 211 is used to extract feature information from the fundus image, and the output networks 212 are used to output the above-described types of information. The fundus image 20 is an image taken by a fundus camera, and includes images of organs such as the macula lutea, optic disc, and blood vessels.
The diabetes type information may be at least two types of information, i.e., information with diabetes or without diabetes, and may further include specific types of diabetes, such as type 1 diabetes and type 2 diabetes; the diabetes-associated disease type information can be at least the presence, absence, or further indicative of a specific category of disease.
According to the method for classifying the diabetes and the related diseases, provided by the embodiment of the invention, the eyeground images of the user are acquired, the characteristics of the eyeground images are extracted by using the machine learning model, and the classification result for representing the diabetes and the related diseases of the user can be output according to the characteristics. The scheme does not need to collect blood samples or other body indexes of a user, only needs to collect fundus images, realizes a noninvasive classification process, and optimizes classification performance due to the introduction of various types of diabetes and related diseases into the model in the training process, so that the output classification result is more accurate. The type of diabetes and related diseases of the subject can be determined in a short time by using electronic equipment such as a computer, a smart phone, a server and the like, and the method has strong convenience and stability; in addition, the scheme does not need to introduce professional medical equipment such as blood sample collecting equipment and the like, and does not need the participation of doctors or professional researchers, so that the cost of diabetes and related disease classification can be reduced, and reliable reference information is provided for the doctors.
In a preferred embodiment, the model further has an output network for outputting the auxiliary information based on the characteristic information. The auxiliary information may be physical indicators or information related to diabetes, or physical indicators or information related to complications or complications of diabetes.
One embodiment of the present invention provides a diabetes classification method, which involves a machine learning model, and the model can be obtained by training using the method described in the above embodiment. The method of the embodiment comprises the following steps: a fundus image of the user is acquired, and as shown in fig. 4, the fundus image 20 is an image taken by a fundus camera, and includes images of organs such as the macula lutea, the optic disc, and blood vessels.
And identifying the fundus images by using a machine learning model and outputting a classification result. The machine learning model in this embodiment includes a feature extraction network 211 and a unique output network 212. Wherein the feature extraction network 211 is used for extracting feature information from the fundus image, and the output network 212 is used for outputting a classification result according to the feature information. The classification result may be information on the type of diabetes, or information on the type of diabetic complication or complication.
In a preferred embodiment, the user may provide one or more types of disease information related to a disease desired to be identified, such as a model for outputting a diabetic nephropathy classification result, the user may provide diabetes type information, cataract type information, and the like as auxiliary information; and one or more easily acquired auxiliary information, such as some non-invasively acquired physical indicators, as auxiliary information. The output network 212 may output the classification results in conjunction with the feature information and the auxiliary information described above.
The model does not have other output networks, does not present other related information during classification, but optimizes the performance according to various types of information and auxiliary information during model training, so that the output classification result is more accurate, the result is concise and intuitive, and the method is suitable for application scenes of rapid evaluation.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It should be understood that the above examples are only for clarity of illustration and are not intended to limit the embodiments. Other variations and modifications will be apparent to persons skilled in the art in light of the above description. And are neither required nor exhaustive of all embodiments. And obvious variations or modifications therefrom are within the scope of the invention.
Claims (11)
1. A method for constructing a classification model of diabetes and related diseases is characterized by comprising the following steps:
acquiring sample data which comprises fundus images and a plurality of types of information, wherein the plurality of types of information are diabetes type information and at least one type of diabetes-related disease information, or at least two types of diabetes-related disease type information;
training a machine learning model by using a large amount of the sample data to output an evaluation result, wherein the evaluation result at least comprises a classification result corresponding to one type of the type information, and the machine learning model comprises a feature extraction network and at least one output network, wherein the feature extraction network is used for extracting feature information from the fundus image, and the at least one output network is used for outputting the evaluation result according to the feature information respectively; and the machine learning model at least adjusts the parameters of the machine learning model according to the output evaluation result and the type information in the sample data.
2. The method according to claim 1, wherein there are a plurality of said output networks, and the output networks respectively output the classification results corresponding to different types of information according to the feature information; the machine learning model adjusts parameters of the machine learning model at least according to the output evaluation result and type information in the sample data, and the method comprises the following steps:
determining a second loss value according to the difference between the classification result and the corresponding type information;
determining a first loss value according to the second loss value;
and the machine learning model adjusts self parameters according to the first loss value.
3. The method according to claim 1, wherein the sample data further comprises auxiliary information, the auxiliary information comprising diabetes related information and/or diabetes associated disease related information.
4. The method according to claim 3, wherein the output network is plural, and includes an output network for outputting classification results corresponding to different types of information according to the feature information, and an output network for outputting recognition results corresponding to the auxiliary information according to the feature information; the evaluation result includes a classification result corresponding to each of the types of information and an identification result corresponding to the auxiliary information;
the machine learning model adjusts parameters of the machine learning model according to the output evaluation result and type information in the sample data, and the method comprises the following steps:
determining a second loss value according to the difference between the classification result and the corresponding type information;
determining a third loss value according to the difference between the auxiliary information and the corresponding recognition result;
determining a first loss value according to the second loss value and the third loss value;
and the machine learning model adjusts self parameters according to the first loss value.
5. The method of claim 2 or 4, wherein the machine learning model adjusts parameters of the feature extraction network based at least on the first loss value.
6. The method of claim 2, wherein the machine learning model adjusts a parameter of the respective output network based on the second loss value.
7. The method of claim 4, wherein the machine learning model adjusts a parameter of the respective output network based on the second loss value and the third loss value.
8. The method of any one of claims 1-7, wherein the diabetes-associated disorder comprises a diabetic complication and/or diabetic complication.
9. A method for classifying diabetes and related disorders, comprising: acquiring a fundus image of a user; identifying the fundus image using a machine learning model constructed by the method of any one of claims 1-8, outputting at least one classification result.
10. A diabetes classification model construction apparatus, comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to cause the at least one processor to perform the method of constructing a classification model for diabetes and related diseases as claimed in any one of claims 1 to 8.
11. An apparatus for classifying diabetes and related conditions, comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the one processor to cause the at least one processor to perform the method of classifying diabetes and related conditions of claim 10.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911421191.8A CN111080643A (en) | 2019-12-31 | 2019-12-31 | Method and device for classifying diabetes and related diseases based on fundus images |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911421191.8A CN111080643A (en) | 2019-12-31 | 2019-12-31 | Method and device for classifying diabetes and related diseases based on fundus images |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111080643A true CN111080643A (en) | 2020-04-28 |
Family
ID=70321327
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911421191.8A Pending CN111080643A (en) | 2019-12-31 | 2019-12-31 | Method and device for classifying diabetes and related diseases based on fundus images |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111080643A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111696100A (en) * | 2020-06-17 | 2020-09-22 | 上海鹰瞳医疗科技有限公司 | Method and device for determining smoking degree based on fundus image |
| CN112686855A (en) * | 2020-12-28 | 2021-04-20 | 博奥生物集团有限公司 | Information correlation method for elephant and symptom information |
| CN112800941A (en) * | 2021-01-26 | 2021-05-14 | 中科人工智能创新技术研究院(青岛)有限公司 | Face anti-fraud method and system based on asymmetric auxiliary information embedded network |
| CN113012148A (en) * | 2021-04-14 | 2021-06-22 | 中国人民解放军总医院第一医学中心 | Diabetic nephropathy-non-diabetic nephropathy differential diagnosis device based on fundus image |
| CN113449781A (en) * | 2021-06-17 | 2021-09-28 | 上海深至信息科技有限公司 | Generation method and system of thyroid nodule classification model |
| CN114255875A (en) * | 2022-02-28 | 2022-03-29 | 北京鹰瞳科技发展股份有限公司 | Methods, devices, systems, and computer-readable storage media for assessing risk of kidney disease |
| WO2022152165A1 (en) * | 2021-01-13 | 2022-07-21 | 高雄医学大学 | Method and system for predicting stage of diabetic nephropathy by means of ophthalmoscopic images |
| CN116433644A (en) * | 2023-04-22 | 2023-07-14 | 深圳市江机实业有限公司 | Eye image dynamic diagnosis method based on recognition model |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108095683A (en) * | 2016-11-11 | 2018-06-01 | 北京羽医甘蓝信息技术有限公司 | The method and apparatus of processing eye fundus image based on deep learning |
| CN109614925A (en) * | 2017-12-07 | 2019-04-12 | 深圳市商汤科技有限公司 | Dress ornament attribute recognition approach and device, electronic equipment, storage medium |
| CN109886143A (en) * | 2019-01-29 | 2019-06-14 | 上海鹰瞳医疗科技有限公司 | Multi-tag disaggregated model training method and equipment |
| CN110135528A (en) * | 2019-06-13 | 2019-08-16 | 上海鹰瞳医疗科技有限公司 | Age determines that method, eye health degree determine method and apparatus |
| CN110210544A (en) * | 2019-05-24 | 2019-09-06 | 上海联影智能医疗科技有限公司 | Image classification method, computer equipment and storage medium |
-
2019
- 2019-12-31 CN CN201911421191.8A patent/CN111080643A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108095683A (en) * | 2016-11-11 | 2018-06-01 | 北京羽医甘蓝信息技术有限公司 | The method and apparatus of processing eye fundus image based on deep learning |
| CN109614925A (en) * | 2017-12-07 | 2019-04-12 | 深圳市商汤科技有限公司 | Dress ornament attribute recognition approach and device, electronic equipment, storage medium |
| CN109886143A (en) * | 2019-01-29 | 2019-06-14 | 上海鹰瞳医疗科技有限公司 | Multi-tag disaggregated model training method and equipment |
| CN110210544A (en) * | 2019-05-24 | 2019-09-06 | 上海联影智能医疗科技有限公司 | Image classification method, computer equipment and storage medium |
| CN110135528A (en) * | 2019-06-13 | 2019-08-16 | 上海鹰瞳医疗科技有限公司 | Age determines that method, eye health degree determine method and apparatus |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111696100A (en) * | 2020-06-17 | 2020-09-22 | 上海鹰瞳医疗科技有限公司 | Method and device for determining smoking degree based on fundus image |
| CN112686855A (en) * | 2020-12-28 | 2021-04-20 | 博奥生物集团有限公司 | Information correlation method for elephant and symptom information |
| CN112686855B (en) * | 2020-12-28 | 2024-04-16 | 博奥生物集团有限公司 | Information association method of eye image and symptom information |
| WO2022152165A1 (en) * | 2021-01-13 | 2022-07-21 | 高雄医学大学 | Method and system for predicting stage of diabetic nephropathy by means of ophthalmoscopic images |
| TWI789199B (en) * | 2021-01-13 | 2023-01-01 | 高雄醫學大學 | Diabetic kidney disease prediction with retinography and system thereof |
| CN112800941A (en) * | 2021-01-26 | 2021-05-14 | 中科人工智能创新技术研究院(青岛)有限公司 | Face anti-fraud method and system based on asymmetric auxiliary information embedded network |
| CN113012148A (en) * | 2021-04-14 | 2021-06-22 | 中国人民解放军总医院第一医学中心 | Diabetic nephropathy-non-diabetic nephropathy differential diagnosis device based on fundus image |
| CN113449781B (en) * | 2021-06-17 | 2023-04-07 | 上海深至信息科技有限公司 | Generation method and system of thyroid nodule classification model |
| CN113449781A (en) * | 2021-06-17 | 2021-09-28 | 上海深至信息科技有限公司 | Generation method and system of thyroid nodule classification model |
| CN114255875B (en) * | 2022-02-28 | 2022-06-21 | 北京鹰瞳科技发展股份有限公司 | Methods, devices, systems, and computer-readable storage media for assessing risk of kidney disease |
| CN114255875A (en) * | 2022-02-28 | 2022-03-29 | 北京鹰瞳科技发展股份有限公司 | Methods, devices, systems, and computer-readable storage media for assessing risk of kidney disease |
| CN116433644A (en) * | 2023-04-22 | 2023-07-14 | 深圳市江机实业有限公司 | Eye image dynamic diagnosis method based on recognition model |
| CN116433644B (en) * | 2023-04-22 | 2024-03-08 | 深圳市江机实业有限公司 | Eye image dynamic diagnosis method based on recognition model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111080643A (en) | Method and device for classifying diabetes and related diseases based on fundus images | |
| Li et al. | Computer‐assisted diagnosis for diabetic retinopathy based on fundus images using deep convolutional neural network | |
| CN108021916B (en) | Deep learning diabetic retinopathy sorting technique based on attention mechanism | |
| WO2022188489A1 (en) | Training method and apparatus for multi-mode multi-disease long-tail distribution ophthalmic disease classification model | |
| Yang et al. | SUD-GAN: deep convolution generative adversarial network combined with short connection and dense block for retinal vessel segmentation | |
| CN111048210B (en) | Method and equipment for evaluating disease risk based on fundus image | |
| CN110197493A (en) | Eye fundus image blood vessel segmentation method | |
| CN111028232A (en) | Diabetes classification method and equipment based on fundus images | |
| CN107423571A (en) | Diabetic retinopathy identifying system based on eye fundus image | |
| CN111696100A (en) | Method and device for determining smoking degree based on fundus image | |
| CN111047590A (en) | Hypertension classification method and device based on fundus images | |
| CN113011340B (en) | Cardiovascular operation index risk classification method and system based on retina image | |
| CN117952964B (en) | Fundus medical image analysis method based on computer vision technology | |
| CN117457229A (en) | Anesthesia depth monitoring system and method based on artificial intelligence | |
| CN112869697A (en) | Judgment method for simultaneously identifying stage and pathological change characteristics of diabetic retinopathy | |
| CN114724231A (en) | Glaucoma multi-modal intelligent recognition system based on transfer learning | |
| CN112052874B (en) | Physiological data classification method and system based on generation countermeasure network | |
| Reddy et al. | Discovering optimal algorithm to predict diabetic retinopathy using novel assessment methods | |
| CN116563932A (en) | Eye image recognition method and related equipment based on multitask learning | |
| Kabir et al. | Multi-classification based Alzheimer's disease detection with comparative analysis from brain MRI scans using deep learning | |
| Leopold et al. | Segmentation and feature extraction of retinal vascular morphology | |
| CN111582366B (en) | Image processing method, device and equipment | |
| CN117672445A (en) | Diabetes mellitus debilitation current situation analysis method and system based on big data | |
| Sudha et al. | Detection and Classification of Diabetic Retinopathy Using DCNN and BSN Models. | |
| CN113576399B (en) | Sugar net analysis method, system and electronic equipment |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20210721 Address after: 100083 room 21, 4th floor, building 2, national defense science and Technology Park, beipolytechnic, Haidian District, Beijing Applicant after: Beijing Yingtong Technology Development Co.,Ltd. Applicant after: SHANGHAI YINGTONG MEDICAL TECHNOLOGY Co.,Ltd. Address before: 200030 room 01, 8 building, 1 Yizhou Road, Xuhui District, Shanghai, 180 Applicant before: SHANGHAI YINGTONG MEDICAL TECHNOLOGY Co.,Ltd. |
|
| TA01 | Transfer of patent application right |