CN112750530A - Model training method, terminal device and storage medium - Google Patents
Model training method, terminal device and storage medium Download PDFInfo
- Publication number
- CN112750530A CN112750530A CN202110009535.5A CN202110009535A CN112750530A CN 112750530 A CN112750530 A CN 112750530A CN 202110009535 A CN202110009535 A CN 202110009535A CN 112750530 A CN112750530 A CN 112750530A
- Authority
- CN
- China
- Prior art keywords
- training
- logistic regression
- focus image
- regression model
- model
- 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
- 238000012549 training Methods 0.000 title claims abstract description 138
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000007477 logistic regression Methods 0.000 claims abstract description 92
- 238000012545 processing Methods 0.000 claims abstract description 35
- 238000011156 evaluation Methods 0.000 claims abstract description 28
- 230000009466 transformation Effects 0.000 claims description 36
- 230000003902 lesion Effects 0.000 claims description 23
- 238000012795 verification Methods 0.000 claims description 17
- 238000007781 pre-processing Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 4
- 238000003745 diagnosis Methods 0.000 abstract description 16
- 238000004891 communication Methods 0.000 description 20
- 201000010099 disease Diseases 0.000 description 15
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 15
- 238000004590 computer program Methods 0.000 description 10
- 230000000694 effects Effects 0.000 description 8
- 238000002591 computed tomography Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 6
- 238000013473 artificial intelligence Methods 0.000 description 4
- 230000001419 dependent effect Effects 0.000 description 4
- 230000003321 amplification Effects 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000002372 labelling Methods 0.000 description 3
- 238000012417 linear regression Methods 0.000 description 3
- 210000004185 liver Anatomy 0.000 description 3
- 238000003199 nucleic acid amplification method Methods 0.000 description 3
- 238000011282 treatment Methods 0.000 description 3
- 238000005034 decoration Methods 0.000 description 2
- 230000008030 elimination Effects 0.000 description 2
- 238000003379 elimination reaction Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000002595 magnetic resonance imaging Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000007170 pathology Effects 0.000 description 2
- 238000013519 translation Methods 0.000 description 2
- 238000002604 ultrasonography Methods 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 235000002566 Capsicum Nutrition 0.000 description 1
- 238000005481 NMR spectroscopy Methods 0.000 description 1
- 239000006002 Pepper Substances 0.000 description 1
- 235000016761 Piper aduncum Nutrition 0.000 description 1
- 235000017804 Piper guineense Nutrition 0.000 description 1
- 244000203593 Piper nigrum Species 0.000 description 1
- 235000008184 Piper nigrum Nutrition 0.000 description 1
- 238000005299 abrasion Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 239000008280 blood Substances 0.000 description 1
- 210000004369 blood Anatomy 0.000 description 1
- 210000004204 blood vessel Anatomy 0.000 description 1
- 210000000988 bone and bone Anatomy 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 201000011510 cancer Diseases 0.000 description 1
- 230000010267 cellular communication Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000003111 delayed effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 238000001839 endoscopy Methods 0.000 description 1
- 201000006674 extrapulmonary tuberculosis Diseases 0.000 description 1
- 230000002496 gastric effect Effects 0.000 description 1
- 210000002216 heart Anatomy 0.000 description 1
- 201000009663 hepatic tuberculosis Diseases 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000008676 import Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 201000007270 liver cancer Diseases 0.000 description 1
- 208000014018 liver neoplasm Diseases 0.000 description 1
- 238000000491 multivariate analysis Methods 0.000 description 1
- 210000003205 muscle Anatomy 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 230000005813 organ abnormality Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000001959 radiotherapy Methods 0.000 description 1
- 150000003839 salts Chemical class 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000003746 surface roughness Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000011426 transformation method Methods 0.000 description 1
- 201000008827 tuberculosis Diseases 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/03—Recognition of patterns in medical or anatomical images
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Public Health (AREA)
- Medical Informatics (AREA)
- Biomedical Technology (AREA)
- Theoretical Computer Science (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Epidemiology (AREA)
- Pathology (AREA)
- Databases & Information Systems (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a model training method, a terminal device and a storage medium, wherein the method comprises the following steps: carrying out image expansion processing on the focus image to obtain a sample set; the sample set comprises a preset number of focus image samples; the focus image sample is provided with an artificial mark; randomly removing a plurality of focus image samples from the sample set, and taking the sample set after the samples are removed as a training set; the number of the focus image samples in the training set is less than the preset number and more than half of the preset number; extracting case characteristics of focus image samples in the training set, and training according to the case characteristics to obtain a target logistic regression model; and evaluating the target logistic regression model, and training according to an evaluation result to obtain a patient identification model. According to the invention, under the condition of a small amount of sample data, a patient identification model with high reliability and high accuracy is obtained through training, the diagnosis accuracy is improved, and the life safety of the patient is improved.
Description
Technical Field
The present invention relates to the field of machine learning technologies, and in particular, to a model training method, a terminal device, and a storage medium.
Background
Currently, more than 70% of the diagnoses in clinic rely on medical images, including ultrasound, pathology, endoscopy, CT (computed tomography), CR (computed tomography), MRI (magnetic resonance imaging), and many other means; the artificial intelligence is mainly applied to medical images by deep learning, realizes the analysis and judgment of a machine on the medical images, assists doctors to finish diagnosis, and is applied to the stages of disease screening, diagnosis and treatment.
Currently, medical image-assisted diagnosis is considered to be one of the most important potential innovative applications of artificial intelligence. The medical imaging doctor has a huge gap, the annual medical image data growth rate of China is about 30%, the annual growth rate of the number of radiologists is only 4%, and the physicians need to study and train for a long time, which means that the imaging physicians have greater and greater pressure on processing image data in the future and are difficult to bear huge load. Meanwhile, with the advance of the graded diagnosis and treatment policy and the release of the basic medical requirements, the medical data will grow faster, and the problem of shortage of radiotherapy department/pathology department doctors will be more serious.
Medical image data is almost manually analyzed by professionals, and the defects of the medical image data are obvious, namely, the diagnosis misdiagnosis rate is high and the diagnosis efficiency is low. According to statistics, the number of misdiagnoses in the United states reaches 1200 ten thousand every year, and the number of misdiagnoses in China reaches 5700 ten thousand every year. According to a misdiagnosis data of the Chinese medical society, the misdiagnosis rate of Chinese clinical medicine is 27.8%, wherein the average misdiagnosis rate of malignant tumors is 40%, the misdiagnosis rate of organ abnormality is 60%, and the average misdiagnosis rate of extrapulmonary tuberculosis such as hepatic tuberculosis, gastric tuberculosis and the like is more than 40%.
Therefore, diagnosis of diseases by using medical images is an auxiliary diagnosis method which is generally applied at present, and although the total amount of data of the medical images is huge, due to the fact that medical diseases are various and the situation of isolated medical information is not well relieved at present, the total amount of samples for training and learning under each type of specific diseases is insufficient, accuracy of a trained model is insufficient, and misdiagnosis is caused, and timely treatment of patients is delayed.
Disclosure of Invention
The invention aims to provide a model training method, terminal equipment and a storage medium, which can be used for training a patient identification model with high reliability and high accuracy under the condition of a small amount of sample data, improving the diagnosis accuracy and improving the life safety of a patient.
The technical scheme provided by the invention is as follows:
the invention provides a model training method, which comprises the following steps:
carrying out image expansion processing on the focus image to obtain a sample set; the sample set comprises a preset number of focus image samples; the focus image sample is provided with an artificial mark;
randomly removing a plurality of focus image samples from the sample set, and taking the sample set with the samples removed as a training set; the number of the focus image samples in the training set is less than the preset number and more than half of the preset number;
extracting case characteristics of focus image samples in the training set, and training according to the case characteristics to obtain a target logistic regression model;
and evaluating the target logistic regression model, and training according to an evaluation result to obtain a patient identification model.
Further, the image expansion processing on the lesion image to obtain the sample set includes the steps of:
and acquiring a focus image after manual marking, and performing transformation processing on the focus image according to a preset image transformation strategy combination to complete image expansion to obtain the sample set.
Further, the extracting of the case characteristics of the focus image samples in the training set and the training according to the case characteristics to obtain the target logistic regression model includes the steps of:
image preprocessing is carried out on each focus image sample in the training set, and image features are extracted from the preprocessed focus image samples and serve as case features;
training according to the case characteristics and preset training parameters to obtain a logistic regression model, and inputting a preset verification set into the logistic regression model to obtain an accuracy value;
if the accuracy value is larger than or equal to a preset value, taking the logistic regression model as a target logistic regression model;
and if the accuracy value is smaller than a preset value, acquiring a new training set for retraining until the target logistic regression model is acquired.
Further, the step of evaluating the target logistic regression model and training to obtain the patient identification model according to the evaluation result comprises the following steps:
drawing a receiver operation characteristic curve according to the target logistic regression model, and calculating to obtain a target area under the receiver operation characteristic curve;
comparing the target area with a preset area threshold value to obtain an evaluation result;
if the evaluation result is that the target area is smaller than or equal to the preset area threshold, determining that the target logistic regression model does not meet the requirement, and acquiring a new training set for retraining;
and if the evaluation result is that the target area is larger than the preset area threshold value, determining the target logistic regression model as the patient identification model.
The present invention also provides a terminal device, including:
the sample expansion module is used for carrying out image expansion processing on the focus image to obtain a sample set; the sample set comprises a preset number of focus image samples; the focus image sample is provided with an artificial mark;
the sample acquisition module is used for randomly eliminating a plurality of focus image samples from the sample set and taking the sample set after the samples are eliminated as a training set; the number of the focus image samples in the training set is less than the preset number and more than half of the preset number;
the model training module is used for extracting case characteristics of the focus image samples in the training set and training according to the case characteristics to obtain a target logistic regression model;
and the model determining module is used for evaluating the target logistic regression model and training according to an evaluation result to obtain a patient identification model.
Further, the sample expansion module comprises:
the acquisition unit is used for acquiring a focus image after artificial marking;
and the expansion unit is used for performing transformation processing on the focus image according to a preset image transformation strategy combination to complete image expansion to obtain the sample set.
Further, the model training module comprises:
the extraction unit is used for carrying out image preprocessing on each focus image sample in the training set and extracting image features from the preprocessed focus image samples to be used as the case features;
the training unit is used for training according to the case characteristics and preset training parameters to obtain a logistic regression model, and inputting a preset verification set into the logistic regression model to obtain an accuracy value;
the first processing unit is used for taking the logistic regression model as a target logistic regression model if the accuracy value is greater than or equal to a preset value; and if the accuracy value is smaller than a preset value, acquiring a new training set for retraining until the target logistic regression model is acquired.
Further, the model determination module comprises:
the calculation unit is used for drawing a receiver operation characteristic curve according to the target logistic regression model and calculating and obtaining a target area under the receiver operation characteristic curve;
the comparison unit is used for comparing the target area with a preset area threshold value to obtain an evaluation result;
the second processing unit is used for determining that the target logistic regression model does not meet the requirements and acquiring a new training set for retraining if the evaluation result is that the target area is smaller than or equal to the preset area threshold; and if the evaluation result is that the target area is larger than the preset area threshold value, determining the target logistic regression model as the patient identification model.
The invention also provides a storage medium, wherein at least one instruction is stored in the storage medium, and the instruction is loaded and executed by a processor to realize the operation executed by the model training method.
By the model training method, the terminal equipment and the storage medium, the patient identification model with high reliability and high accuracy can be obtained by training under the condition of a small amount of sample data, the diagnosis accuracy is improved, and the life safety of the patient is improved.
Drawings
The above features, technical features, advantages and implementations of a model training method, terminal device and storage medium will be further described in the following detailed description of preferred embodiments in a clearly understandable manner, with reference to the accompanying drawings.
FIG. 1 is a flow chart of one embodiment of a model training method of the present invention;
FIG. 2 is a flow chart of another embodiment of a model training method of the present invention.
Detailed Description
In the following description, for purposes of explanation and not limitation, specific details are set forth, such as particular system structures, techniques, etc. in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to one skilled in the art that the present application may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.
It will be understood that the terms "comprises" and/or "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
For the sake of simplicity, the drawings only schematically show the parts relevant to the present invention, and they do not represent the actual structure as a product. In addition, in order to make the drawings concise and understandable, components having the same structure or function in some of the drawings are only schematically illustrated or only labeled. In this document, "one" means not only "only one" but also a case of "more than one".
It should be further understood that the term "and/or" as used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.
In addition, in the description of the present application, the terms "first", "second", and the like are used only for distinguishing the description, and are not intended to indicate or imply relative importance.
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following description will be made with reference to the accompanying drawings. It is obvious that the drawings in the following description are only some examples of the invention, and that for a person skilled in the art, other drawings and embodiments can be derived from them without inventive effort.
In an embodiment of the present invention, as shown in fig. 1, a method for training a model includes:
s100, carrying out image expansion processing on the focus image to obtain a sample set; the sample set comprises a preset number of focus image samples; the focus image sample is provided with an artificial mark;
specifically, the problem often encountered by artificial intelligence or machine learning is that the data volume of the training set is not enough, or the training set has data loss (i.e. the universal variables are missing, the records are incomplete, and are inconsistent and therefore unavailable, etc.), which often results in the accuracy of the trained patient recognition model being not enough. The invention aims to solve the problems of insufficient data quantity and data loss of a training set of an artificial intelligence algorithm for disease risk prediction, and the specific process is as follows:
a lesion image, such as a skin CT image, a liver CT image, a head X-ray image, a heart, brain, blood vessel, bone, muscle nuclear magnetic resonance image, or an organ B-ultrasound image, is acquired from a hospital side or a patient side. The disease examination report provided to the patient by the hospital can be a paper report or an electronic report. Therefore, if the disease is a paper report, the focus image is image data including the paper report, which is obtained by shooting through a camera after the focus image covers or blocks preset privacy information (including contact information, contact addresses, identification numbers and the like) of a user. Of course, if the electronic report is provided, the lesion image is the image data of the electronic report acquired by the screen capture software, and then the processed image data acquired by the preset privacy information (including contact information, contact address, identification number, etc.) of the user is covered or shielded by the software processing method.
After the preset number of focus images are obtained in the above manner, professional medical staff identify the current focus image in a manual labeling manner to judge whether the user has the corresponding disease, if so, manually label the current focus image as "diseased", otherwise, manually label the current focus image as "non-diseased", and so on, and manually label the preset number of focus images according to the above manner until all the obtained focus images are labeled. After the preset number of focus images are obtained in the above manner, sample amplification is performed on the preset number of focus images by using an image expansion technology to obtain a sample set.
For example, assuming that a patient identification model for determining whether the patient has the liver cancer needs to be trained, 100 different liver CT images can be obtained, and the 100 different liver CT images are labeled manually by the above method.
S200, randomly removing a plurality of focus image samples from the sample set, and taking the sample set with the removed samples as a training set; the number of the focus image samples in the training set is less than the preset number and more than half of the preset number;
s300, extracting case characteristics of focus image samples in the training set, and training according to the case characteristics to obtain a target logistic regression model;
s400, evaluating the target logistic regression model, and training according to an evaluation result to obtain a patient identification model.
Specifically, after the sample set is obtained in the above manner, a plurality of focus image samples are removed from the sample set in a random selection manner, and then the sample set from which the plurality of focus image samples are removed is used as a training set, and it should be noted that the number of focus image samples in the training set is less than the preset number and is greater than half of the preset number. Therefore, after the training set is obtained, image feature extraction is carried out on the remaining focus image samples in the training set, and then case features corresponding to the focus image samples in the training set are extracted and obtained. And training according to the case characteristics to obtain a target logistic regression model, evaluating the target logistic regression model, and training according to an evaluation result to obtain a patient identification model.
According to the invention, a large number of focus images are obtained through amplification by carrying out image expansion processing on a small number of focus images to obtain a sample set, then the focus image samples in the sample set are subjected to random elimination processing to obtain a training set, model training is carried out based on the training set, the condition of data loss can be simulated, and then a patient identification model with high reliability and high accuracy can be trained under the condition of a small number of sample data, so that the diagnosis accuracy is improved, and the life safety of a patient is improved.
In an embodiment of the present invention, as shown in fig. 2, a method for training a model includes:
s110, acquiring a focus image after artificial labeling, and performing transformation processing on the focus image according to a preset image transformation strategy combination to complete image expansion to obtain the sample set;
specifically, the preset image transformation strategy combination comprises a physical transformation and a superposition transformation, wherein the physical transformation comprises any one or more combinations of a translation transformation, a rotation transformation, a turning transformation, a scaling transformation, a contrast transformation, a brightness transformation, a fuzzy transformation, a noise detour transformation and a scene effect superposition transformation. And performing corresponding transformation processing on all the generated focus images according to any image transformation mode in the preset image transformation strategy combination to generate new focus images, and completing image expansion to obtain a sample set.
Wherein the translation transform is: and translating the focus image left and right, up and down to an appointed distance or a random distance. The rotation transformation is: and rotating the focus image by taking the midpoint as a reference at a designated or random angle. The flip transformation is: the lesion image is flipped around the horizontal or vertical axis at a designated or random angle. The scaling transform is: and (3) amplifying or reducing the focus image at a specified or random ratio. The contrast transformation is: the color tone is unchanged, and the saturation and brightness of the focus image are changed. The luminance transform is: the brightness is unchanged, and the saturation and the color tone of the focus image are changed. Fuzzy transformation: and assigning all pixel points in the focus image as pixel average values to perform fuzzy processing. The noise perturbation is transformed into: noise (e.g., gaussian noise and salt and pepper noise) is used to add noise to the lesion image. The scene effect superposition transform is: and adding watermark icons of the effects of disease examination report cracks, disease examination report shelters, disease examination report dirty stains (including blood stains, mud stains, oil stains, water stains and the like), disease examination report abrasion and the like on the focus images.
Illustratively, after the focus image is transformed by the physical transformation method, a small number of candidate transformation images are randomly selected from the transformation images, and a watermark icon of a disease examination report water stain effect is added to the candidate transformation images to generate a new focus image. Other preset image transformation strategy combinations are within the protection scope of the invention, and are not described in detail herein.
S200, randomly removing a plurality of focus image samples from the sample set, and taking the sample set with the removed samples as a training set; the number of the focus image samples in the training set is less than the preset number and more than half of the preset number;
s310, image preprocessing is carried out on each focus image sample in the training set, and image features are extracted from the preprocessed focus image samples and serve as case features;
specifically, image preprocessing is performed on all focus image samples in the training set by image processing methods such as gray level processing and binarization processing, image recognition is performed on the preprocessed focus image samples, and image features corresponding to the preprocessed focus image samples are extracted as case features, wherein the image features comprise shapes, contours, sizes, positions, surface roughness and the like of lesion areas.
S320, training according to the case characteristics and preset training parameters to obtain a logistic regression model, and inputting a preset verification set into the logistic regression model to obtain an accuracy value;
specifically, the logistic regression (logistic regression) model is a multivariate analysis method for studying the relationship between the dependent variable and the influence factor (independent variable) and belongs to probabilistic nonlinear regression. Logistic regression, also called generalized linear regression model, is basically the same form as the linear regression model, both having ax + b, where a and b are parameters to be solved, and is distinguished by their dependent variables, the multiple linear regression directly takes ax + b as the dependent variable, i.e. y ═ ax + b, while logistic regression has the functional form p ═ locations wTx + b, where w is the weight value when the cost function is minimized, and b is the offset when the cost function is minimized. The value of the dependent variable is then determined based on the size of p and 1-p. After the case features are extracted through the embodiment, training parameters are initialized and set, wherein the training parameters comprise w, b and a cost function. Generally, w is randomly initialized to a value close to 0, and b is initialized to 0. Then, a logistic regression model is obtained through training according to the case characteristics and the training parameters. And storing the parameters a and b trained by the model so as to directly import the trained parameters during prediction.
S330, if the accuracy value is larger than or equal to a preset value, taking the logistic regression model as a target logistic regression model;
s340, if the accuracy value is smaller than a preset value, acquiring a new training set for retraining until the target logistic regression model is acquired;
specifically, after the sample set is obtained by the embodiment, all the focus image samples in the sample set are obtained into the training set according to the above manner, and the randomly selected focus image samples are used as the verification set. It should be noted that the division of the training set and the validation set is to keep the data distribution as consistent as possible. The sample number of the focus image samples in the training set and the verification set is divided according to the proportion of 7:3 as much as possible, so that the efficiency and the accuracy of subsequent clinical auxiliary model training are improved. Of course, it is within the scope of the present invention to divide the lesion image samples according to other ratios, as long as the number of the samples of the lesion image samples in the training set is greater than the number of the samples of the lesion image samples in the verification set.
And after the verification set is obtained by division, inputting the focus image samples in the verification set into a logistic regression model obtained by training to obtain an accuracy value, and if the accuracy value is greater than or equal to a preset value, taking the logistic regression model obtained by current training as a target logistic regression model. Of course, if the accuracy value is smaller than the preset value, obtaining a new training set to retrain in the manner of the above embodiment, and determining to obtain the target logistic regression model until the accuracy of the logistic regression model obtained by inputting the lesion image samples in the newly divided verification set to retrain is greater than or equal to the preset value.
S410, drawing a receiver operation characteristic curve according to the target logistic regression model, and calculating to obtain a target area under the receiver operation characteristic curve;
s420, comparing the target area with a preset area threshold value to obtain an evaluation result;
s430, if the evaluation result is that the target area is smaller than or equal to the preset area threshold, determining that the target logistic regression model does not meet the requirement, and acquiring a new training set for retraining;
s440, if the evaluation result indicates that the target area is larger than the preset area threshold, determining that the target logistic regression model is the patient identification model.
Specifically, after the target logistic regression model is obtained in the above embodiments, a true positive TP, a true negative TN, a false negative FN and a false positive FP may be obtained, where TP is the number of samples that belong to a lesion after the lesion image samples marked as a definite lesion in the verification set are judged by using the target logistic regression model, TN is the number of samples that belong to a lesion after the lesion image samples marked as a definite lesion in the verification set are judged by using the target logistic regression model, FN is the number of samples that belong to a lesion after the lesion image samples marked as a definite lesion in the verification set are judged by using the target logistic regression model, and FP is the number of samples that belong to a lesion after the lesion image samples marked as a definite lesion in the verification set are judged by using the target logistic regression model. Then, calculating a pseudo-positive rate FPR and a real case rate TPR according to TP, TN, FN and FP, drawing an ROC curve (namely a receiver operation characteristic curve of the invention) by taking the FPR as an abscissa and the TPR as an ordinate, calculating an AUC area (namely a target area of the invention) under the ROC curve, comparing the AUC area with a preset area threshold value so as to evaluate the target logistic regression model, judging that the target logistic regression model is not qualified when the AUC area is less than or equal to the preset area threshold value, returning to the initial step to obtain a new training set, and retraining until the target logistic regression model meeting the requirement is obtained. Of course, if the AUC area is greater than the preset area threshold, it is determined that the target logistic regression model meets the requirement, and the target logistic regression model meeting the requirement is determined as the patient identification model.
According to the embodiment of the invention, the target logistic regression model is trained through an iterative algorithm, the ROC curve is drawn according to the target logistic regression model, the target logistic regression model is evaluated according to the AUC area under the ROC curve, and the patient identification model is obtained through training. Through the mode, the verification set is screened in a manual labeling mode, so that the classification accuracy of the focus image samples can be improved, the identification accuracy of a patient identification model is improved, meanwhile, a target logistic regression model meeting the requirements is screened out by adopting the area AUC under the ROC curve to serve as the patient identification model, the identification accuracy of whether the focus is the focus can be further improved, the primary diagnosis of a doctor on the disease can be subsequently performed, important references are provided for the disease diagnosis, the diagnosis accuracy and efficiency are effectively improved, and the workload of the doctor is greatly reduced.
According to the invention, a large number of focus images are obtained through amplification by carrying out image expansion processing on a small number of focus images to obtain a sample set, then the focus image samples in the sample set are subjected to random elimination processing to obtain a training set, model training is carried out based on the training set, the condition of data loss can be simulated, and then a patient identification model with high reliability and high accuracy can be trained under the condition of a small number of sample data, so that the diagnosis accuracy is improved, and the life safety of a patient is improved.
In one embodiment of the present invention, a terminal device includes:
the sample expansion module is used for carrying out image expansion processing on the focus image to obtain a sample set; the sample set comprises a preset number of focus image samples; the focus image sample is provided with an artificial mark;
the sample acquisition module is used for randomly eliminating a plurality of focus image samples from the sample set and taking the sample set after the samples are eliminated as a training set; the number of the focus image samples in the training set is less than the preset number and more than half of the preset number;
the model training module is used for extracting case characteristics of the focus image samples in the training set and training according to the case characteristics to obtain a target logistic regression model;
and the model determining module is used for evaluating the target logistic regression model and training according to an evaluation result to obtain a patient identification model.
Specifically, this embodiment is a device embodiment corresponding to the method embodiment, and specific effects refer to the method embodiment, which is not described in detail herein.
Based on the foregoing embodiments, the sample expansion module includes:
the acquisition unit is used for acquiring a focus image after artificial marking;
and the expansion unit is used for performing transformation processing on the focus image according to a preset image transformation strategy combination to complete image expansion to obtain the sample set.
Specifically, this embodiment is a device embodiment corresponding to the method embodiment, and specific effects refer to the method embodiment, which is not described in detail herein.
Based on the foregoing embodiments, the model training module includes:
the extraction unit is used for carrying out image preprocessing on each focus image sample in the training set and extracting image features from the preprocessed focus image samples to be used as the case features;
the training unit is used for training according to the case characteristics and preset training parameters to obtain a logistic regression model, and inputting a preset verification set into the logistic regression model to obtain an accuracy value;
the first processing unit is used for taking the logistic regression model as a target logistic regression model if the accuracy value is greater than or equal to a preset value; and if the accuracy value is smaller than a preset value, acquiring a new training set for retraining until the target logistic regression model is acquired.
Specifically, this embodiment is a device embodiment corresponding to the method embodiment, and specific effects refer to the method embodiment, which is not described in detail herein.
Based on the foregoing embodiments, the model determination module includes:
the calculation unit is used for drawing a receiver operation characteristic curve according to the target logistic regression model and calculating and obtaining a target area under the receiver operation characteristic curve;
the comparison unit is used for comparing the target area with a preset area threshold value to obtain an evaluation result;
the second processing unit is used for determining that the target logistic regression model does not meet the requirements and acquiring a new training set for retraining if the evaluation result is that the target area is smaller than or equal to the preset area threshold; and if the evaluation result is that the target area is larger than the preset area threshold value, determining the target logistic regression model as the patient identification model.
Specifically, this embodiment is a device embodiment corresponding to the method embodiment, and specific effects refer to the method embodiment, which is not described in detail herein.
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-described division of program modules is illustrated, and in practical applications, the above-described distribution of functions may be performed by different program modules, that is, the internal structure of the apparatus may be divided into different program units or modules to perform all or part of the above-described functions. Each program module in the embodiments may be integrated in one processing unit, or each unit may exist alone physically, or two or more units are integrated in one processing unit, and the integrated unit may be implemented in a form of hardware, or may be implemented in a form of software program unit. In addition, the specific names of the program modules are only used for distinguishing the program modules from one another, and are not used for limiting the protection scope of the application.
In an embodiment of the present invention, a terminal device 100 includes a processor 110, a memory 120, wherein the memory 120 is used for storing a computer program 121; the processor 110 is configured to execute the computer program 121 stored in the memory 120 to implement the training method of the model in the corresponding method embodiment.
The terminal device 100 may be a desktop computer, a notebook, a palm computer, a tablet computer, a mobile phone, a human-computer interaction screen, or the like. The terminal device 100 may include, but is not limited to, a processor 110, a memory 120. Those skilled in the art will appreciate that the foregoing is merely an example of the terminal device 100 and does not constitute a limitation of the terminal device 100 and may include more or less components than those shown, or some of the components may be combined, or different components, such as: the terminal device 100 may also include input/output interfaces, display devices, network access devices, communication buses, communication interfaces, and the like. A communication interface and a communication bus, and may further include an input/output interface, wherein the processor 110, the memory 120, the input/output interface and the communication interface complete communication with each other through the communication bus. The memory 120 stores a computer program 121, and the processor 110 is configured to execute the computer program 121 stored in the memory 120 to implement the model training method in the corresponding method embodiment.
The Processor 110 may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, discrete hardware component, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The memory 120 may be an internal storage unit of the terminal device 100, such as: hard disk or memory of the terminal device. The memory may also be an external storage medium of the terminal device, such as: the terminal equipment is provided with a plug-in hard disk, an intelligent memory Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card) and the like. Further, the memory 120 may also include both an internal storage unit and an external storage medium of the terminal device 100. The memory 120 is used for storing the computer program 121 and other programs and data required by the terminal device 100. The memory may also be used to temporarily store data that has been output or is to be output.
A communication bus is a circuit that connects the described elements and enables transmission between the elements. For example, the processor 110 receives commands from other elements through the communication bus, decrypts the received commands, and performs calculations or data processing according to the decrypted commands. The memory 120 may include program modules such as a kernel (kernel), middleware (middleware), an Application Programming Interface (API), and applications. The program modules may be comprised of software, firmware or hardware, or at least two of the same. The input/output interface forwards commands or data entered by a user via the input/output interface (e.g., sensor, keyboard, touch screen). The communication interface connects the terminal device 100 with other network devices, user equipment, networks. For example, the communication interface may be connected to a network by wire or wirelessly to connect to external other network devices or user devices. The wireless communication may include at least one of: wireless fidelity (WiFi), Bluetooth (BT), Near Field Communication (NFC), Global Positioning Satellite (GPS) and cellular communications, among others. The wired communication may include at least one of: universal Serial Bus (USB), high-definition multimedia interface (HDMI), asynchronous transfer standard interface (RS-232), and the like. The network may be a telecommunications network and a communications network. The communication network may be a computer network, the internet of things, a telephone network. The terminal device 100 may be connected to the network through a communication interface, and a protocol by which the terminal device 100 communicates with other network devices may be supported by at least one of an application, an Application Programming Interface (API), middleware, a kernel, and a communication interface.
In an embodiment of the present invention, a storage medium stores at least one instruction, and the instruction is loaded and executed by a processor to implement the operations performed by the corresponding embodiments of the training method for the model. For example, the storage medium may be a read-only memory (ROM), a Random Access Memory (RAM), a compact disc read-only memory (CD-ROM), a magnetic tape, a floppy disk, an optical data storage medium, and the like.
They may be implemented in program code that is executable by a computing device such that it is executed by the computing device, or separately, or as individual integrated circuit modules, or as a plurality or steps of individual integrated circuit modules. Thus, the present invention is not limited to any specific combination of hardware and software.
In the above embodiments, the descriptions of the respective embodiments have respective emphasis, and reference may be made to the related descriptions of other embodiments for parts that are not described or recited in detail in a certain embodiment.
Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus/terminal device and method may be implemented in other ways. For example, the above-described embodiments of the apparatus/terminal device are merely illustrative, and for example, the division of the modules or units is only one logical division, and there may be other divisions when actually implemented, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection of some interfaces, devices or units, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated modules/units may be stored in a storage medium if they are implemented in the form of software functional units and sold or used as separate products. Based on such understanding, all or part of the flow in the method according to the embodiments of the present invention may also be implemented by sending instructions to relevant hardware by the computer program 121, where the computer program 121 may be stored in a storage medium, and when the computer program 121 is executed by a processor, the steps of the above-described embodiments of the method may be implemented. The computer program 121 may be in a source code form, an object code form, an executable file or some intermediate form, etc. The storage medium may include: any entity or device capable of carrying the computer program 121, recording medium, usb disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM), Random Access Memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the content of the storage medium may be increased or decreased as appropriate according to the requirements of legislation and patent practice in the jurisdiction, for example: in certain jurisdictions, in accordance with legislation and patent practice, computer-readable storage media do not include electrical carrier signals and telecommunications signals.
It should be noted that the above embodiments can be freely combined as necessary. The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.
Claims (9)
1. A method of training a model, comprising the steps of:
carrying out image expansion processing on the focus image to obtain a sample set; the sample set comprises a preset number of focus image samples; the focus image sample is provided with an artificial mark;
randomly removing a plurality of focus image samples from the sample set, and taking the sample set with the samples removed as a training set; the number of the focus image samples in the training set is less than the preset number and more than half of the preset number;
extracting case characteristics of focus image samples in the training set, and training according to the case characteristics to obtain a target logistic regression model;
and evaluating the target logistic regression model, and training according to an evaluation result to obtain a patient identification model.
2. The method for training the model according to claim 1, wherein the step of performing image expansion processing on the lesion image to obtain the sample set comprises the steps of:
and acquiring the artificially marked focus image, and performing transformation processing on the focus image according to a preset image transformation strategy combination to complete image expansion to obtain the sample set.
3. The method for training the model according to claim 1, wherein the step of extracting the case characteristics of the focus image samples in the training set and training the case characteristics to obtain the target logistic regression model comprises the steps of:
image preprocessing is carried out on each focus image sample in the training set, and image features are extracted from the preprocessed focus image samples and serve as case features;
training according to the case characteristics and preset training parameters to obtain a logistic regression model, and inputting a preset verification set into the logistic regression model to obtain an accuracy value;
if the accuracy value is larger than or equal to a preset value, taking the logistic regression model as a target logistic regression model;
and if the accuracy value is smaller than a preset value, acquiring a new training set for retraining until the target logistic regression model is acquired.
4. A method for training a model according to any one of claims 1-3, wherein the step of evaluating the target logistic regression model and training a patient identification model according to the evaluation result comprises the steps of:
drawing a receiver operation characteristic curve according to the target logistic regression model, and calculating to obtain a target area under the receiver operation characteristic curve;
comparing the target area with a preset area threshold value to obtain an evaluation result;
if the evaluation result is that the target area is smaller than or equal to the preset area threshold, determining that the target logistic regression model does not meet the requirement, and acquiring a new training set for retraining;
and if the evaluation result is that the target area is larger than the preset area threshold value, determining the target logistic regression model as the patient identification model.
5. A terminal device, comprising:
the sample expansion module is used for carrying out image expansion processing on the focus image to obtain a sample set; the sample set comprises a preset number of focus image samples; the focus image sample is provided with an artificial mark;
the sample acquisition module is used for randomly eliminating a plurality of focus image samples from the sample set and taking the sample set after the samples are eliminated as a training set; the number of the focus image samples in the training set is less than the preset number and more than half of the preset number;
the model training module is used for extracting case characteristics of the focus image samples in the training set and training according to the case characteristics to obtain a target logistic regression model;
and the model determining module is used for evaluating the target logistic regression model and training according to an evaluation result to obtain a patient identification model.
6. The terminal device of claim 5, wherein the sample expansion module comprises:
the acquisition unit is used for acquiring a focus image after artificial marking;
and the expansion unit is used for performing transformation processing on the focus image according to a preset image transformation strategy combination to complete image expansion to obtain the sample set.
7. The terminal device of claim 5, wherein the model training module comprises:
the extraction unit is used for carrying out image preprocessing on each focus image sample in the training set and extracting image features from the preprocessed focus image samples to be used as the case features;
the training unit is used for training according to the case characteristics and preset training parameters to obtain a logistic regression model, and inputting a preset verification set into the logistic regression model to obtain an accuracy value;
the first processing unit is used for taking the logistic regression model as a target logistic regression model if the accuracy value is greater than or equal to a preset value; and if the accuracy value is smaller than a preset value, acquiring a new training set for retraining until the target logistic regression model is acquired.
8. The terminal device of any of claims 5-7, wherein the model determination module comprises:
the calculation unit is used for drawing a receiver operation characteristic curve according to the target logistic regression model and calculating and obtaining a target area under the receiver operation characteristic curve;
the comparison unit is used for comparing the target area with a preset area threshold value to obtain an evaluation result;
the second processing unit is used for determining that the target logistic regression model does not meet the requirements and acquiring a new training set for retraining if the evaluation result is that the target area is smaller than or equal to the preset area threshold; and if the evaluation result is that the target area is larger than the preset area threshold value, determining the target logistic regression model as the patient identification model.
9. A storage medium having stored therein at least one instruction which is loaded and executed by a processor to perform operations performed by a training method of a model according to any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110009535.5A CN112750530A (en) | 2021-01-05 | 2021-01-05 | Model training method, terminal device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110009535.5A CN112750530A (en) | 2021-01-05 | 2021-01-05 | Model training method, terminal device and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112750530A true CN112750530A (en) | 2021-05-04 |
Family
ID=75649873
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110009535.5A Pending CN112750530A (en) | 2021-01-05 | 2021-01-05 | Model training method, terminal device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112750530A (en) |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103488889A (en) * | 2013-09-18 | 2014-01-01 | 刘华锋 | Method and system of detecting artifacts in ICU patient records based on multivariate logistic regression |
| CN107273429A (en) * | 2017-05-19 | 2017-10-20 | 哈工大大数据产业有限公司 | A kind of Missing Data Filling method and system based on deep learning |
| CN107704495A (en) * | 2017-08-25 | 2018-02-16 | 平安科技（深圳）有限公司 | Training method, device and the computer-readable recording medium of subject classification device |
| CN109146899A (en) * | 2018-08-28 | 2019-01-04 | 众安信息技术服务有限公司 | CT image jeopardizes organ segmentation method and device |
| CN109584218A (en) * | 2018-11-15 | 2019-04-05 | 首都医科大学附属北京友谊医院 | A kind of construction method of gastric cancer image recognition model and its application |
| CN110197236A (en) * | 2018-10-17 | 2019-09-03 | 中山大学附属第三医院 | A kind of prediction and verification method based on image group to glucocorticoid curative effect |
| CN110837572A (en) * | 2019-11-15 | 2020-02-25 | 北京推想科技有限公司 | Image retrieval method and device, readable storage medium and electronic equipment |
| US20200126662A1 (en) * | 2017-02-20 | 2020-04-23 | Ping An Technology (Shenzhen) Co., Ltd. | Method, device, and apparatus for detecting disease probability, and computer-readable storage medium |
| CN111370059A (en) * | 2020-03-05 | 2020-07-03 | 上海市肺科医院（上海市职业病防治院） | Prediction system and method for tumor mutation load |
| CN111403024A (en) * | 2019-01-02 | 2020-07-10 | 天津幸福生命科技有限公司 | Method and device for obtaining disease judgment model based on medical data |
| CN111430025A (en) * | 2020-03-10 | 2020-07-17 | 清华大学 | Disease diagnosis method based on medical image data amplification |
| CN112132833A (en) * | 2020-08-25 | 2020-12-25 | 沈阳工业大学 | Skin disease image focus segmentation method based on deep convolutional neural network |
| CN112150429A (en) * | 2020-09-18 | 2020-12-29 | 南京师范大学 | Attention mechanism guided kidney CT image segmentation method |
-
2021
- 2021-01-05 CN CN202110009535.5A patent/CN112750530A/en active Pending
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103488889A (en) * | 2013-09-18 | 2014-01-01 | 刘华锋 | Method and system of detecting artifacts in ICU patient records based on multivariate logistic regression |
| US20200126662A1 (en) * | 2017-02-20 | 2020-04-23 | Ping An Technology (Shenzhen) Co., Ltd. | Method, device, and apparatus for detecting disease probability, and computer-readable storage medium |
| CN107273429A (en) * | 2017-05-19 | 2017-10-20 | 哈工大大数据产业有限公司 | A kind of Missing Data Filling method and system based on deep learning |
| CN107704495A (en) * | 2017-08-25 | 2018-02-16 | 平安科技（深圳）有限公司 | Training method, device and the computer-readable recording medium of subject classification device |
| CN109146899A (en) * | 2018-08-28 | 2019-01-04 | 众安信息技术服务有限公司 | CT image jeopardizes organ segmentation method and device |
| CN110197236A (en) * | 2018-10-17 | 2019-09-03 | 中山大学附属第三医院 | A kind of prediction and verification method based on image group to glucocorticoid curative effect |
| CN109584218A (en) * | 2018-11-15 | 2019-04-05 | 首都医科大学附属北京友谊医院 | A kind of construction method of gastric cancer image recognition model and its application |
| CN111403024A (en) * | 2019-01-02 | 2020-07-10 | 天津幸福生命科技有限公司 | Method and device for obtaining disease judgment model based on medical data |
| CN110837572A (en) * | 2019-11-15 | 2020-02-25 | 北京推想科技有限公司 | Image retrieval method and device, readable storage medium and electronic equipment |
| CN111370059A (en) * | 2020-03-05 | 2020-07-03 | 上海市肺科医院（上海市职业病防治院） | Prediction system and method for tumor mutation load |
| CN111430025A (en) * | 2020-03-10 | 2020-07-17 | 清华大学 | Disease diagnosis method based on medical image data amplification |
| CN112132833A (en) * | 2020-08-25 | 2020-12-25 | 沈阳工业大学 | Skin disease image focus segmentation method based on deep convolutional neural network |
| CN112150429A (en) * | 2020-09-18 | 2020-12-29 | 南京师范大学 | Attention mechanism guided kidney CT image segmentation method |
Non-Patent Citations (8)
| Title |
|---|
| 付等: "基于深度学习的阴道镜及磁共振图像子宫识别研究", 《中国医学计算机成像杂志》 * |
| 刘鹏等: "面向高维缺失数据集的线性判别分析方法", 《常州大学学报(自然科学版)》 * |
| 孙聚波等: "基于E-MS算法的含缺失数据图模型选择的模拟研究", 《佳木斯大学学报(自然科学版)》 * |
| 张力等: "CT影像组学在头颈部木村病淋巴结病变与淋巴瘤鉴别中的应用", 《放射学实践》 * |
| 方晨晨等: "基于改进深度残差网络的番茄病害图像识别", 《计算机应用》 * |
| 汪娟等: "基于改进的GLCM甲状腺纹理特征提取与分析", 《计算机工程与应用》 * |
| 石祥滨等: "小样本高光谱遥感图像深度学习方法", 《系统仿真学报》 * |
| 蒋佳旺等: "卷积神经网络与迁移学习的颅脑癌症识别方法的研究", 《中国医疗设备》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108464840B (en) | Automatic detection method and system for breast lumps | |
| CN109949271B (en) | Detection method based on medical image, model training method and device | |
| Rangan et al. | Quantification of renal pathology by image analysis (Methods in Renal Research) | |
| US8335359B2 (en) | Systems, apparatus and processes for automated medical image segmentation | |
| Isasi et al. | Melanomas non-invasive diagnosis application based on the ABCD rule and pattern recognition image processing algorithms | |
| US9858665B2 (en) | Medical imaging device rendering predictive prostate cancer visualizations using quantitative multiparametric MRI models | |
| WO2021135774A1 (en) | Tumor prediction method and device, cloud platform, and computer-readable storage medium | |
| Garcia-Hernandez et al. | Analysis of the impact of digital watermarking on computer-aided diagnosis in medical imaging | |
| KR102600401B1 (en) | Apparatus, method and computer program for anayzing medical image using classification and segmentation | |
| WO2019146358A1 (en) | Learning system, method, and program | |
| CN112288843A (en) | Three-dimensional construction method and device of focus, terminal device and storage medium | |
| CN112071418B (en) | Gastric cancer peritoneal metastasis prediction system and method based on enhanced CT image histology | |
| CN112634231A (en) | Image classification method and device, terminal equipment and storage medium | |
| CN114332132A (en) | Image segmentation method and device and computer equipment | |
| Alsheh Ali et al. | Detection of potential microcalcification clusters using multivendor for‐presentation digital mammograms for short‐term breast cancer risk estimation | |
| Rosado et al. | 12 From Dermoscopy to Mobile Teledermatology | |
| CN115274094A (en) | Liver cancer imaging combined AFP diagnosis model construction method, system and device | |
| Kumar et al. | A Novel Approach for Breast Cancer Detection by Mammograms | |
| CN113130050A (en) | Medical information display method and system | |
| Albiol et al. | Automatic intensity windowing of mammographic images based on a perceptual metric | |
| CN116912247A (en) | Medical image processing method and device, storage medium and electronic equipment | |
| CN112750530A (en) | Model training method, terminal device and storage medium | |
| CN115841476A (en) | Method, device, equipment and medium for predicting life cycle of liver cancer patient | |
| CN113822917A (en) | Accurate registration method for liver cancer imaging omics images | |
| CN115578285B (en) | Mammary gland molybdenum target image detail enhancement method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210504 |