CN113160978A - Full femtosecond postoperative vision prediction method, system and medium based on machine learning - Google Patents
Full femtosecond postoperative vision prediction method, system and medium based on machine learning Download PDFInfo
- Publication number
- CN113160978A CN113160978A CN202011444808.0A CN202011444808A CN113160978A CN 113160978 A CN113160978 A CN 113160978A CN 202011444808 A CN202011444808 A CN 202011444808A CN 113160978 A CN113160978 A CN 113160978A
- Authority
- CN
- China
- Prior art keywords
- vision
- module
- prediction
- full femtosecond
- postoperative
- 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
- 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
- G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Data Mining & Analysis (AREA)
- Biomedical Technology (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Epidemiology (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Eye Examination Apparatus (AREA)
Abstract
The invention discloses a full femtosecond postoperative vision prediction method based on machine learning, which comprises the following steps: acquiring a full femtosecond operation sample data set; performing effective variable screening on the full femtosecond operation sample data set; inputting an AI algorithm to the effective variable for model training; predicting postoperative vision of the trained mathematical model; carrying out error adjustment on the prediction result; and converting the vision expression form of the adjusted prediction result. The invention discloses a full femtosecond postoperative vision prediction system based on machine learning, which comprises a full femtosecond operation sample data set acquisition module, an effective variable screening module, an AI algorithm model training module, an postoperative vision prediction module, a prediction result error adjustment module and a vision expression form conversion module. The invention can accurately predict the vision conditions of the early and the long term after the full femtosecond operation.
Description
Technical Field
The invention relates to the technical field of artificial intelligence for postoperative vision prediction of ophthalmology, in particular to a full femtosecond postoperative vision prediction method, a system and a medium based on machine learning.
Background
Since 2008, femtosecond laser Small-incision stromal lens extraction (called "full femtosecond" for short) has been widely recommended and has become a great breakthrough in "minimally invasive surgery" in corneal refractive surgery. This surgical approach avoids the occurrence of corneal flaps and all of the corneal flap-related complications. To date, there have been over 200 million worldwide SMILE procedures. Moreover, increasing clinical studies have demonstrated that SMILE is a safe, accurate, and effective refractive procedure. SMILE has many advantages over other types of corneal refractive surgery, such as: iatrogenic dry eye has a low incidence, faster recovery of corneal perception, and biomechanical advantages. However, in some cases, refractive power regression and loss of naked eye vision occurred after SMILE surgery. Naked eye vision directly represents the ability of a patient to identify objects without wearing corrective glasses. Therefore, the method is very significant for predicting the naked eye vision state of the patient after SMILE operation.
In the past traditional studies, the results of visual status after SMILE surgery are reported generally as "signature groups". Thus, under practical clinical scenarios, the clinician can only make a rough estimate of the post-SMILE vision status "after having a rich clinical experience and learning more clinical study reports". Admittedly, even the estimation of such a situation still requires a considerable expenditure of time, effort and corresponding effort on the part of the clinician. In fact, the information that each patient possesses is a "big data" challenge for the clinician. Computers undoubtedly go far beyond the human brain in terms of their ability to process large amounts of information and storage. Therefore, in contrast to relying solely on brain memory analysis, computer Machine Learning (ML), i.e., Artificial Intelligence (AI), can help perform predictive analysis, thereby assisting clinicians in making proper clinical decisions. Currently, AI has been applied maturely to simulate the learning results of humans. AI has made substantial progress in ophthalmology, particularly with respect to the convenience of image recognition. This will, of course, advance the leap in the technical qualities of ophthalmic AI. Undeniably, there is currently a lack of accurate prediction of post-SMILE vision. The post-operative vision status is critical to both the SMILE surgeon and the patient who is about to receive SMILE for myopia correction for the following reasons: first, accurate prediction of post-SMILE vision can help reduce psychological stress on SMILE patients and surgeons; secondly, accurate prediction of post-SMILE vision is more helpful for determining the indication of SMILE; in addition, conventional research reports and techniques do not meet these needs.
Disclosure of Invention
In view of the above, there is a need to provide a full femtosecond postoperative vision prediction method, system and medium based on machine learning, which can accurately predict the early and long-term vision conditions after SMILE operation.
In order to achieve the purpose, the invention adopts the following technical scheme:
the invention provides a full femtosecond postoperative vision prediction method based on machine learning, which comprises the following steps:
s1, acquiring a full femtosecond operation sample data set;
s2, performing effective variable screening on the full femtosecond operation sample data set obtained in the S1;
s3, inputting an AI algorithm to the effective variables obtained in S2 for model training;
s4, predicting postoperative vision of the mathematical model trained in the S3;
s5, carrying out error adjustment on the prediction result obtained in the S4;
and S6, converting the vision expression form of the prediction result after the adjustment of the S5.
Further, in S1, the full femtosecond surgical sample dataset includes biological information, a refractive surgical preoperative examination parameter set, a surgical parameter set, and a post-operative obtained parameter set; the refractive pre-operation examination parameter group comprises an optimal correction vision value, a myopic sphere degree, an astigmatic cylinder degree, an equivalent sphere degree, a thinnest point corneal thickness, a flat corneal curvature, a steep corneal curvature, an intraocular pressure value and a normal range value corresponding to each parameter in a plurality of groups; the surgery parameter group comprises an additional spherical lens power, a corneal cap diameter, a corneal cap thickness, a lens light area diameter, a lens thickness, a residual matrix thickness and a normal range value corresponding to each parameter in a plurality of groups; the parameter group obtained after the operation comprises negative pressure suction time, an naked eye vision value for a plurality of days after the operation, an naked eye vision value for a plurality of years after the operation and a normal range value corresponding to each parameter in a plurality of groups.
Further, S2 includes the steps of:
s21, performing correlation analysis on parameters in the full femtosecond operation sample data set to identify variables in a correlation coefficient range;
s22, the variable obtained in S21 is ranked in importance degree by the Boruta algorithm to distinguish the variable within an important score range, so as to obtain a valid variable.
Further, S22 includes the steps of:
s221, judging the average reduction precision after the characteristic change according to the importance of the variable, evaluating the importance of each characteristic, namely the variable, and iteratively and gradually deleting the non-important characteristics;
and S223, obtaining the result of accepting or rejecting the characteristic variable so as to realize model variable screening.
Further, the AI algorithm is an extremely random tree algorithm.
Further, in S4, the trained mathematical model is subjected to vision prediction several days after the operation and several years after the operation.
Further, in S6, the prediction result is converted into a visual expression form, and the logMAR expression form, the Decimal expression form, or the Line expression form is obtained after the conversion.
The invention provides a full femtosecond postoperative vision prediction system based on machine learning, which comprises a full femtosecond operation sample data set acquisition module, an effective variable screening module, an AI algorithm model training module, an postoperative vision prediction module, a prediction result error adjustment module and a vision expression form conversion module;
the full femtosecond operation sample data set acquisition module is used for acquiring a full femtosecond operation sample data set;
the effective variable screening module is used for carrying out effective variable screening on the full femtosecond operation sample data set obtained from the full femtosecond operation sample data set acquisition module;
the AI algorithm model training module is used for inputting the effective variables obtained from the effective variable screening module into an AI algorithm for model training;
the postoperative vision prediction module is used for predicting postoperative vision of the trained mathematical model obtained from the AI algorithm model training module;
the prediction result error adjusting module is used for carrying out error adjustment on the prediction result obtained from the postoperative vision predicting module;
and the vision expression form conversion module is used for converting the vision expression form of the prediction result obtained by the prediction result error adjustment module after error adjustment.
The invention proposes a computer-readable storage medium, on which a computer program is stored, which, when executed by a processor, implements a machine learning-based all-femtosecond postoperative vision prediction method as defined in any one of the above.
The invention has the beneficial effects that:
compared with the prior art, the full-femtosecond postoperative vision prediction method, system and medium based on machine learning disclosed by the invention have the advantages that relevant preoperative and estimated surgical parameters acquired clinically in the prior art are subjected to parameter screening, screened effective variables are input into a full-femtosecond vision prediction AI algorithm model, and the prediction vision after SMILE is provided. The doctor can predict the vision after the full femtosecond operation according to the result output by the SMLE postoperative vision prediction AI algorithm model, help the clinician and the patient to better evaluate the self demand and the communication before the operation, and solve the regret that the SMLE postoperative vision prediction is lacked in the prior art; the method comprises the steps of preprocessing clinical detection parameters in the prior art to obtain effective prediction parameters, inputting the effective prediction parameters into a full femtosecond vision prediction AI algorithm model to obtain early and long-term prediction vision after SMILE operation, and directly predicting the vision state after the full femtosecond operation by a doctor through the full femtosecond vision prediction AI algorithm model, thereby reducing the uncontrollable risk of the doctor in work and being beneficial to the selection of full femtosecond adaptation diseases; in addition, the AI algorithm model obtained by training a large number of real world parameters with determined clinical results is high in prediction accuracy, so that the situation of possible prediction errors caused by insufficient clinical experience of doctors in practice is avoided, and the blind area of the technical problem in the prior art is filled.
Drawings
Fig. 1 is a working flow chart of a full femtosecond postoperative vision prediction method based on machine learning.
Fig. 2 is a conversion table of the vision expression form according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be further clearly and completely described below with reference to the embodiments of the present invention. It should be noted that the described embodiments are only a part of the embodiments of the present invention, and not all embodiments. 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.
Examples
As shown in fig. 1, the invention provides a machine learning-based all-femtosecond postoperative vision prediction method, which comprises the following steps:
s1, acquiring a full femtosecond operation sample data set;
s2, performing effective variable screening on the full femtosecond operation sample data set obtained in the S1;
s3, inputting an AI algorithm to the effective variables obtained in S2 for model training;
s4, predicting postoperative vision of the mathematical model trained in the S3;
s5, carrying out error adjustment on the prediction result obtained in the S4;
and S6, converting the vision expression form of the prediction result after the adjustment of the S5.
Specifically, the method can provide parameters, screen and predict relevant effective variables, establish and evaluate an AI algorithm model for predicting the early vision after the full femtosecond operation, establish and evaluate an AI model for predicting the long-term vision after the SMILE operation and provide vision prediction values and conversion of vision expression forms to realize accurate prediction of the vision after the full femtosecond operation by collecting and preparing inputs.
Further, in S1, the full femtosecond surgical sample dataset includes biological information, a refractive surgical preoperative examination parameter set, a surgical parameter set, and a post-operative obtained parameter set; the refractive pre-operation examination parameter group comprises an optimal correction vision value, a myopic sphere degree, an astigmatic cylinder degree, an equivalent sphere degree, a thinnest point corneal thickness, a flat corneal curvature, a steep corneal curvature, an intraocular pressure value and a normal range value corresponding to each parameter in a plurality of groups; the surgery parameter group comprises an additional spherical lens power, a corneal cap diameter, a corneal cap thickness, a lens light area diameter, a lens thickness, a residual matrix thickness and a normal range value corresponding to each parameter in a plurality of groups; the parameter group obtained after the operation comprises negative pressure suction time, an naked eye vision value for a plurality of days after the operation, an naked eye vision value for a plurality of years after the operation and a normal range value corresponding to each parameter in a plurality of groups.
Specifically, in S1, a full femtosecond surgical sample dataset is acquired, for predicting potentially relevant parameters, see table 1 below:
TABLE 1
Further, S2 includes the steps of:
s21, performing correlation analysis on parameters in the full femtosecond operation sample data set to identify variables in a correlation coefficient range; specifically, strongly correlated variables are identified by correlation analysis of the parameters;
s22, sorting the importance degrees of the variables obtained in the S21 through a Boruta algorithm to distinguish the variables in an important score range, so as to obtain effective variables; specifically, the specific steps of screening the variables by adopting the Boruta algorithm are that the average reduction precision after the characteristic change is judged according to the importance of database variable signs, namely the variables, so as to evaluate the importance of each characteristic, namely the variable, the non-important characteristics are iteratively and gradually deleted, and finally, the result of accepting or rejecting the characteristic variables is given, so that the purpose of screening the model variables is achieved; and selecting the variables with low relevance and importance, namely the screened effective variables.
In one of the screening examples, the final effective variables were determined as follows:
TABLE 2
Specifically, in S2, the performing effective variable screening on the acquired full femtosecond vision prediction data set includes: effective variable screening, namely feature selection, is a process of selecting some most effective features from a group of features to reduce the spatial dimension of the features, and is a key step of model establishment. A good training sample is crucial to the classifier, and directly influences the robustness and generalization capability of model prediction.
Further, S22 includes the steps of:
s221, judging the average reduction precision (namely database variable sign) after the characteristic change according to the importance of the variable, evaluating the importance of each characteristic, namely the variable, and iteratively and gradually deleting the non-important characteristic;
and S223, obtaining the result of accepting or rejecting the characteristic variable so as to realize model variable screening.
Further, the AI algorithm is an extreme random tree algorithm; preferably, the Root Mean Square Error (RMSE) and the Mean Absolute Error (MAE) are selected as indexes for evaluating the model effect, and finally the model established by the extratres extremely random tree algorithm is confirmed as the optimal prediction model; extratress is a combinatorial approach to decision trees; similar to random forests, the same bootstrap sample randomly selects partial features to construct a tree, but compared with the random forests, the extremely random trees are more random in selection of division points; compared with a standard random forest algorithm, the extreme random tree makes the decision boundary smoother; the extreme random tree directly constructs a random tree by using the training sample, corrects the bagging mode, and is superior to a random forest in performance when the data noise is large or the data volume is large.
Specifically, in S3, the screened effective variables are input to an AI algorithm for model training, and through 10-fold cross validation, the finally selected AI algorithm model of the present invention is evaluated, so that a plurality of currently approved and recommended AI algorithms (including Lasso, Random Relationships (RF), explicit Random oriented Trees (extra Trees), Gradient Boost Machine (GBM) and explicit Gradient Boosting (XGBoost) algorithms) can be applied for optimization of algorithm parameters, training and evaluation of the model.
Further, in S4, the trained mathematical model is subjected to vision prediction several days after the operation and several years after the operation.
Further, in S6, the prediction result is transformed into a vision expression form, and the transformed logMAR expression form or Decimal expression form or Line expression form (as shown in fig. 2) is obtained, so as to give a specific value of the vision at the early stage or the far stage after the full femtosecond operation.
The invention provides a full femtosecond postoperative vision prediction system based on machine learning, which comprises a full femtosecond operation sample data set acquisition module, an effective variable screening module, an AI algorithm model training module, an postoperative vision prediction module, a prediction result error adjustment module and a vision expression form conversion module;
the full femtosecond operation sample data set acquisition module is used for acquiring a full femtosecond operation sample data set;
the effective variable screening module is used for carrying out effective variable screening on the full femtosecond operation sample data set obtained from the full femtosecond operation sample data set acquisition module;
the AI algorithm model training module is used for inputting the effective variables obtained from the effective variable screening module into an AI algorithm for model training;
the postoperative vision prediction module is used for predicting postoperative vision of the trained mathematical model obtained from the AI algorithm model training module;
the prediction result error adjusting module is used for carrying out error adjustment on the prediction result obtained from the postoperative vision predicting module;
and the vision expression form conversion module is used for converting the vision expression form of the prediction result obtained by the prediction result error adjustment module after error adjustment.
The invention proposes a computer-readable storage medium, on which a computer program is stored, which, when executed by a processor, implements a machine learning-based all-femtosecond postoperative vision prediction method as defined in any one of the above.
The working principle of the invention is as follows:
the invention obtains effective prediction variables by characteristic screening of the parameters to be distinguished, inputs the effective variables into an AI algorithm model for predicting early vision and long-term vision after SMILE operation, adjusts the predicted vision according to error values, and provides conversion of vision expression forms, thereby realizing accurate prediction of early and long-term vision conditions after SMLE operation.
The above-mentioned embodiments only express several embodiments of the present invention, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (9)
1. A full femtosecond postoperative vision prediction method based on machine learning is characterized by comprising the following steps:
s1, acquiring a full femtosecond operation sample data set;
s2, performing effective variable screening on the full femtosecond operation sample data set obtained in the S1;
s3, inputting an AI algorithm to the effective variables obtained in S2 for model training;
s4, predicting postoperative vision of the mathematical model trained in the S3;
s5, carrying out error adjustment on the prediction result obtained in the S4;
and S6, converting the vision expression form of the prediction result after the adjustment of the S5.
2. The method of claim 1, wherein in step S1, the full femtosecond surgical sample data set includes biological information, a set of refractive surgical pre-operative examination parameters, a set of surgical parameters, and a set of post-operative parameters; the refractive pre-operation examination parameter group comprises an optimal correction vision value, a myopic sphere degree, an astigmatic cylinder degree, an equivalent sphere degree, a thinnest point corneal thickness, a flat corneal curvature, a steep corneal curvature, an intraocular pressure value and a normal range value corresponding to each parameter in a plurality of groups; the surgery parameter group comprises an additional spherical lens power, a corneal cap diameter, a corneal cap thickness, a lens light area diameter, a lens thickness, a residual matrix thickness and a normal range value corresponding to each parameter in a plurality of groups; the parameter group obtained after the operation comprises negative pressure suction time, an naked eye vision value for a plurality of days after the operation, an naked eye vision value for a plurality of years after the operation and a normal range value corresponding to each parameter in a plurality of groups.
3. The full femtosecond postoperative vision prediction method based on machine learning according to claim 1, wherein S2 includes the following steps:
s21, performing correlation analysis on parameters in the full femtosecond operation sample data set to identify variables in a correlation coefficient range;
s22, the variable obtained in S21 is ranked in importance degree by the Boruta algorithm to distinguish the variable within an important score range, so as to obtain a valid variable.
4. The full femtosecond postoperative vision prediction method based on machine learning according to claim 3, wherein S22 includes the following steps:
s221, judging the average reduction precision after the characteristic change according to the importance of the variable, evaluating the importance of each characteristic, namely the variable, and iteratively and gradually deleting the non-important characteristics;
and S223, obtaining the result of accepting or rejecting the characteristic variable so as to realize model variable screening.
5. The machine learning-based full femtosecond postoperative vision prediction method according to claim 1, wherein the AI algorithm is an extreme random tree algorithm.
6. The full femtosecond postoperative vision prediction method based on machine learning as claimed in claim 1, wherein in S4, vision prediction is performed on the trained mathematical model after several days of operation and several years of operation.
7. The method for predicting vision after all femtoseconds operation based on machine learning as claimed in claim 1, wherein in S6, the prediction result is transformed into vision expression form to obtain logMAR expression form or Decimal expression form or Line expression form.
8. A full femtosecond postoperative vision prediction system based on machine learning is characterized by comprising a full femtosecond operation sample data set acquisition module, an effective variable screening module, an AI algorithm model training module, a postoperative vision prediction module, a prediction result error adjustment module and a vision expression form conversion module;
the full femtosecond operation sample data set acquisition module is used for acquiring a full femtosecond operation sample data set;
the effective variable screening module is used for carrying out effective variable screening on the full femtosecond operation sample data set obtained from the full femtosecond operation sample data set acquisition module;
the AI algorithm model training module is used for inputting the effective variables obtained from the effective variable screening module into an AI algorithm for model training;
the postoperative vision prediction module is used for predicting postoperative vision of the trained mathematical model obtained from the AI algorithm model training module;
the prediction result error adjusting module is used for carrying out error adjustment on the prediction result obtained from the postoperative vision predicting module;
and the vision expression form conversion module is used for converting the vision expression form of the prediction result obtained by the prediction result error adjustment module after error adjustment.
9. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, implements the machine learning-based all-femtosecond post-operative vision prediction method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011444808.0A CN113160978A (en) | 2020-12-11 | 2020-12-11 | Full femtosecond postoperative vision prediction method, system and medium based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011444808.0A CN113160978A (en) | 2020-12-11 | 2020-12-11 | Full femtosecond postoperative vision prediction method, system and medium based on machine learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113160978A true CN113160978A (en) | 2021-07-23 |
Family
ID=76882533
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011444808.0A Pending CN113160978A (en) | 2020-12-11 | 2020-12-11 | Full femtosecond postoperative vision prediction method, system and medium based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113160978A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114927230A (en) * | 2022-04-11 | 2022-08-19 | 四川大学华西医院 | Machine learning-based severe heart failure patient prognosis decision support system and method |
| CN115985515A (en) * | 2023-03-20 | 2023-04-18 | 广东工业大学 | Amblyopia correction effect prediction method, device and equipment based on machine learning |
-
2020
- 2020-12-11 CN CN202011444808.0A patent/CN113160978A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114927230A (en) * | 2022-04-11 | 2022-08-19 | 四川大学华西医院 | Machine learning-based severe heart failure patient prognosis decision support system and method |
| CN114927230B (en) * | 2022-04-11 | 2023-05-23 | 四川大学华西医院 | Prognosis decision support system and method for severe heart failure patient based on machine learning |
| CN115985515A (en) * | 2023-03-20 | 2023-04-18 | 广东工业大学 | Amblyopia correction effect prediction method, device and equipment based on machine learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107358605B (en) | The deep neural network apparatus and system of diabetic retinopathy for identification | |
| CN109829894A (en) | Parted pattern training method, OCT image dividing method, device, equipment and medium | |
| CN109101994B (en) | Fundus image screening method and device, electronic equipment and storage medium | |
| CN108553079A (en) | Lesion identifying system based on eye fundus image | |
| CN112133441B (en) | Method and terminal for establishing MH postoperative crack state prediction model | |
| CN113160978A (en) | Full femtosecond postoperative vision prediction method, system and medium based on machine learning | |
| CN108335757B (en) | Method for predicting diopter adjustment value in SMILE operation | |
| Lopes et al. | Artificial intelligence in corneal diagnosis: where are we? | |
| CN111414972B (en) | XGBoost-based eye refraction correction multisource data automatic classification method | |
| CN113689954A (en) | Hypertension risk prediction method, device, equipment and medium | |
| CN116433644B (en) | Eye image dynamic diagnosis method based on recognition model | |
| CN113243887B (en) | Intelligent diagnosis and treatment instrument for macular degeneration of old people | |
| Dupps Jr et al. | A large-scale computational analysis of corneal structural response and ectasia risk in myopic laser refractive surgery (An American ophthalmological society thesis) | |
| CN111563884A (en) | Neural network-based fundus disease identification method, computer device, and medium | |
| Tahvildari et al. | Application of artificial intelligence in the diagnosis and management of corneal diseases | |
| US11881310B2 (en) | Intraocular lenses that improve post-surgical spectacle independent and methods of manufacturing thereof | |
| Anand et al. | Visual outcome following cataract surgery in rural Punjab | |
| Choi et al. | Establishment of a prediction tool for ocular trauma patients with machine learning algorithm | |
| CN112036448B (en) | Keratoconus identification method and device | |
| KR102537470B1 (en) | Diagnosis method for Graves’orbitopathy using orbital computed tomography (CT) based on neural network-based algorithm | |
| CN108888407B (en) | Method for predicting diopter adjustment value in SMILE operation based on Softmax regression algorithm | |
| CN115240857A (en) | Artificial crystal prediction method, device, equipment and storage medium | |
| CN115170492A (en) | Intelligent prediction and evaluation system for postoperative vision of cataract patient based on AI (artificial intelligence) technology | |
| CN111951219B (en) | Thyroid eye disease screening method, system and equipment based on orbit CT image | |
| CN111583248B (en) | Processing method based on eye ultrasonic image |
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 |