CN112085830A - Optical coherent angiography imaging method based on machine learning - Google Patents
Optical coherent angiography imaging method based on machine learning Download PDFInfo
- Publication number
- CN112085830A CN112085830A CN201910513946.0A CN201910513946A CN112085830A CN 112085830 A CN112085830 A CN 112085830A CN 201910513946 A CN201910513946 A CN 201910513946A CN 112085830 A CN112085830 A CN 112085830A
- Authority
- CN
- China
- Prior art keywords
- machine learning
- oct
- network model
- training
- images
- 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.)
- Granted
Links
- 238000010801 machine learning Methods 0.000 title claims abstract description 69
- 238000003384 imaging method Methods 0.000 title claims abstract description 33
- 230000003287 optical effect Effects 0.000 title claims abstract description 16
- 238000002583 angiography Methods 0.000 title claims abstract description 15
- 230000001427 coherent effect Effects 0.000 title claims abstract description 8
- 238000012549 training Methods 0.000 claims abstract description 59
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 22
- 238000000034 method Methods 0.000 claims abstract description 16
- 230000000694 effects Effects 0.000 claims abstract description 8
- FCKYPQBAHLOOJQ-UHFFFAOYSA-N Cyclohexane-1,2-diaminetetraacetic acid Chemical compound OC(=O)CN(CC(O)=O)C1CCCCC1N(CC(O)=O)CC(O)=O FCKYPQBAHLOOJQ-UHFFFAOYSA-N 0.000 claims abstract 11
- 238000012360 testing method Methods 0.000 claims description 11
- 238000013528 artificial neural network Methods 0.000 claims description 7
- 238000013527 convolutional neural network Methods 0.000 claims description 7
- 238000011056 performance test Methods 0.000 claims description 7
- 230000003044 adaptive effect Effects 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 230000000306 recurrent effect Effects 0.000 claims description 2
- 238000001514 detection method Methods 0.000 abstract description 2
- 238000002601 radiography Methods 0.000 abstract 2
- 230000002792 vascular Effects 0.000 abstract 1
- 238000012014 optical coherence tomography Methods 0.000 description 62
- 238000005516 engineering process Methods 0.000 description 10
- 210000004204 blood vessel Anatomy 0.000 description 4
- 230000008859 change Effects 0.000 description 4
- 238000013135 deep learning Methods 0.000 description 4
- 238000011161 development Methods 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 210000001525 retina Anatomy 0.000 description 3
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 210000003743 erythrocyte Anatomy 0.000 description 2
- 230000029058 respiratory gaseous exchange Effects 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 210000003484 anatomy Anatomy 0.000 description 1
- 230000003042 antagnostic effect Effects 0.000 description 1
- 238000000149 argon plasma sintering Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000017531 blood circulation Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000019771 cognition Effects 0.000 description 1
- 230000001149 cognitive effect Effects 0.000 description 1
- 238000002591 computed tomography Methods 0.000 description 1
- 239000002872 contrast media Substances 0.000 description 1
- 239000008358 core component Substances 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 230000002496 gastric effect Effects 0.000 description 1
- 238000000338 in vitro Methods 0.000 description 1
- 238000011423 initialization method Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000035479 physiological effects, processes and functions Effects 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012827 research and development Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
- G06T7/337—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving reference images or patches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
- G06T7/344—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10116—X-ray image
- G06T2207/10121—Fluoroscopy
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30101—Blood vessel; Artery; Vein; Vascular
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an optical coherent angiography imaging method based on machine learning. According to the invention, an original data set required by network model training is generated by utilizing OCT three-dimensional structural images of a sample acquired by OCTA equipment, a whole group of OCT structural images with poor registration effect are removed, an OCTA algorithm is adopted for radiography and imaging to generate a training data set, a machine learning network model is established and trained, so that OCTA radiography is carried out through the machine learning network model; the method can play a great role in the field of OCTA, can generate an angiogram with higher signal-to-noise ratio and better vascular connectivity, and inhibits the common speckle effect in OCT images to a great extent; the label image is automatically generated by an algorithm, so that the applicability of the method is expanded without being influenced by system errors caused by different systems; the damage can be reduced by using smaller detection power for imaging, or the data volume required by imaging is reduced during imaging, and scanning can be completed more quickly.
Description
Technical Field
The invention relates to an optical coherent angiography imaging technology, in particular to an optical coherent angiography imaging method based on machine learning.
Background
Optical Coherence Tomography (OCT) is a high-resolution, non-contact, fast three-dimensional imaging technique. The method utilizes the coherence principle of scattered light in biological tissues, and the signal contrast of the method is derived from the difference of light scattering capacities of different biological tissues. The OCT technology combines a semiconductor technology and an ultrafast laser technology, core components such as a broadband light source, a Michelson interferometer, a photoelectric detector and the like are used for obtaining a backscattering signal of biological tissues, and finally a real-time micron-sized tomographic image of the biological tissues can be obtained through digital signal processing of a computer. Therefore, the OCT technique has been one of important means for diagnosing anatomical structures for a long time, and plays an important role not only in ophthalmic clinical examinations but also in fields such as dermatology, gastrointestinal medicine, cardiology, and neurology.
With the development of scientific technology, the OCT technology has experienced many major breakthroughs and developments in software and hardware in the last 30 years, and has faster imaging speed and higher system sensitivity. Particularly, after 2002, with the maturity of the frequency domain OCT technology, the OCT technology has attracted attention and applications in various fields.
In 1991, Huang et al, the American college of Science and technology, built a first OCT prototype with a longitudinal resolution of 15 μm, and published a first in vitro human retina OCT scan image and a corresponding tissue section image in the journal of Science, which verified the feasibility of the OCT system. Wojtkowski et al obtained the first live human eye retina image based on frequency domain OCT technology in 2002, and Johannes and Leitgeb compared frequency domain OCT in various parameters theoretically and experimentally in turn, and proved that frequency domain OCT has higher sensitivity and faster imaging speed. Since then, frequency domain OCT is gradually replacing time domain OCT and has gained wide attention and applications.
Optical coherence Angiography (OCTA) is a new non-invasive vessel imaging technique that has emerged in recent years. During specific imaging, signal light scans a sample through a galvanometer system, a scanning area is generally rectangular and is divided into a fast axis direction and a slow axis direction, and during scanning, the signal light continuously and repeatedly scans for multiple times (generally 4 times) in the fast axis direction so as to record OCT signals of the same position at different moments, then tissue information is removed through algorithm processing, blood flow signals are extracted, and an angiographic image is generated. It skillfully utilizes flowing red blood cells as a contrast agent, namely, when the red blood cells continuously flow in a blood vessel, OCT signals in the blood vessel continuously change, so as to be distinguished from stable signals of static tissues.
The current imaging algorithms of the OCTA are mainly divided into three types of imaging algorithms based on phase change, amplitude change and combined change of phase and amplitude according to the source of blood vessel information. The nature of the method is to compare OCT signals at the same position and different moments in an analytical calculation mode. However, these methods often only use a part of information in the OCT signal, resulting in problems such as low signal-to-noise ratio of the contrast image and severe speckle. The current main means for solving the problem is to increase the number of scans at the same position and enhance the intensity of blood vessel signals, which results in too long scanning time and artifacts caused by the vibration of the sample, such as shaking and breathing of the eyes of a patient during an ophthalmic examination. In addition, long-term laser irradiation can also cause damage to biological tissue.
Disclosure of Invention
In view of the above problems in the prior art, the present invention provides an optical coherence angiography imaging method based on machine learning.
The invention relates to an optical coherent angiography imaging method based on machine learning, which comprises the following steps:
1) generating an original data set:
acquiring an OCT three-dimensional structural image of a sample by using OCTA equipment, and generating an original data set required by network model training, wherein the original data set comprises j multiplied by k groups of OCT structural image sequences, each group of OCT structural image sequences comprises i OCT structural images with two-dimensional cross sections (B-Scan), k is the number of the sample, j is the number of scanning positions of a slow axis of each sample, i is the scanning times of the same scanning position of the slow axis of the same sample, i is a natural number larger than 4, j is a natural number larger than 50, and k is a natural number larger than 5;
2) and (3) screening data:
registering i B-Scan face OCT structural images in the same group of OCT structural images by adopting a rigid registration algorithm, calculating registration accuracy by utilizing a correlation algorithm after registration, removing the whole group of OCT structural images with poor registration effect, and reserving n groups of screened OCT structural images, wherein n is a natural number, and
3) generating a training data set:
carrying out contrast imaging by using the n groups of screened OCT structural images obtained in the step 2) by adopting an OCTA algorithm, wherein each group of OCT structural images obtains a B-Scan surface OCTA contrast image called a label image; taking m B-Scan surface OCT structural images from i B-Scan surface OCT structural images of a group of OCT structural images corresponding to each label image, wherein m is 2,3 or 4, the m is called input data, the input data is matched with the label images, and the input data and the label images form a training data set required by network model training;
4) establishing a machine learning network model:
constructing a machine learning network model, and setting hyper-parameters of the machine learning network model; and dividing the training data set into n1Group training set and n2Group test set, training set and test set independent of each other, n1And n2Are respectively natural numbers, and
5) training a machine learning network model:
using the machine learning network model established in the step 4) to take the input data in the training data set as the machine learningLearning the input of a network model, where n is1The group training set is used for training the machine learning network model and takes n2The group test set is used for verifying the performance of the machine learning network model; in the training process, the training set is divided into a plurality of batches and repeatedly input into the machine learning network model for training for a plurality of times, the difference between the output image and the label image of the machine learning network model is judged or calculated as a training error to train the machine learning network model, after the training of each batch is finished, the test set is used for carrying out performance test on the machine learning network model, when the performance test indexes of the standby machine learning network model tend to be stable, the training of the machine learning network model is considered to be finished, and the trained machine learning network model is stored;
6) performing OCTA imaging by using a machine learning network model:
and (3) by utilizing the trained machine learning network model, taking the OCT structural image of the sample acquired by the OCTA equipment as input, wherein the output image is the OCTA contrast image.
In step 1), when the OCTA device collects a sample, scanning one slow axis scanning position of one sample once to obtain an OCT structural image of one B-Scan plane, scanning the same slow axis scanning position i times, wherein each sample has j slow axis scanning positions and has k samples in total, thereby obtaining OCT structural images of i × j × k B-Scan planes, dividing the OCT structural images of i × j × k B-Scan planes into j × k groups of OCT structural images, and each group of OCT structural images includes i OCT structural images of B-Scan planes, that is, each slow axis scanning position corresponds to one group of OCT structural images, thereby obtaining an original data set.
In step 2), for j × k groups of OCT structural images after registration, the registration accuracy is compared, the first n groups of OCT structural images with higher registration accuracy are retained, and the remaining OCT structural images are considered to have poorer registration effect, and these groups of OCT structural images are rejected.
In the step 4), the machine learning network model adopts a deep convolutional neural network CNN, a generative antagonistic network GAN or a recurrent neural network RNN; the hyper-parameters comprise the number of network layers, convolution kernels, learning rate, parameter initialization, training rounds and batch size.
In step 5), a mean square error, a structural similarity or a peak signal-to-noise ratio between the output image and the label image is used as a training error and a performance test index, and one of a Stochastic Gradient Descent (SGD), an Adaptive Moment Estimation optimization algorithm (Adam) and a Momentum algorithm (Momentum) is used to minimize the training error so as to train the machine learning network model.
In step 6), the OCT structural image of the sample is scanned for a plurality of times at the same slow axis scanning position, and the scanning times are less than or equal to 4.
Machine learning is a necessary product of artificial intelligence research and development to a certain stage, and has become a core research topic of artificial intelligence. The goal is to let the computer obtain knowledge or skills by mimicking the behavior of human learning and to constantly learn new knowledge to improve performance. The machine learning refers to the subjects of physiology, psychology, cognition and the like, and establishes a calculation model or a cognitive model similar to human learning by understanding the self-learning mechanism of human, thereby forming various learning theories and learning methods, and establishing a learning system with specific application facing to specific tasks.
Common algorithms in machine learning at present comprise artificial neural networks, support vector machines, naive Bayes, random forests, sparse dictionaries, reinforcement learning, characterization learning, similarity measurement learning and the like. With the development of computer hardware, deep learning develops gradually, and the deep learning is a comprehensive evolution of an artificial neural network. The depth learning expands the depth and the width of the artificial neural network, and can infinitely approximate a more complex nonlinear model, thereby learning objective rules and internal relations hidden in data. In a general sense, deep learning algorithms include deep belief networks, deep neural networks, and convolutional neural networks, where the deep belief networks and deep neural networks are very similar in structure. The deep learning networks currently used in image processing are convolutional neural networks and generative confrontation networks. Machine learning is widely applied to the fields of reconstruction, enhancement and segmentation of medical images at present, but is not applied to image reconstruction of OCTA (optical computed tomography).
The invention has the advantages that:
the invention can play a great role in the field of OCTA, and the strong data mining capability of the invention helps OCTA equipment to generate an angiogram with higher signal-to-noise ratio and better vessel connectivity, and inhibits the common speckle effect in OCT images to a great extent; it is worth mentioning that the label image in the invention is automatically generated by an algorithm, different from common machine learning application, the label data needs to be obtained by expert labeling, and the applicability of the method is expanded without being influenced by system errors caused by different systems. In addition, in the same OCTA equipment, in order to obtain an OCTA image of the same level, the invention can use smaller detection power to image, reduce the damage of laser to biological tissues (such as ophthalmology), or reduce the data amount required by imaging during imaging, namely reduce the scanning times of the same position, can complete scanning more quickly, and reduce artifacts (such as shaking of a patient, breathing and the like) caused by the sample vibration due to overlong scanning time.
Drawings
FIG. 1 is a flow chart of a machine learning based optical coherence angiography imaging method of the present invention;
fig. 2 is an OCTA image obtained by one embodiment of the machine-learning-based optical coherence angiography imaging method according to the present invention.
Detailed Description
The invention will be further elucidated by means of specific embodiments in the following with reference to the drawing.
The optical coherence angiography imaging method based on machine learning of the embodiment, as shown in fig. 1, includes the following steps:
1) generating an original data set:
the method comprises the steps that an OCTA device is used for collecting obtained OCT three-dimensional structural images of a retina to generate an original data set required by network model training, the same slow axis scanning position of the same sample is scanned for 50 times, the slow axis scanning position of each sample is 100, the samples are 30 human eyes, and therefore the generated original data set comprises 100 x 30 groups of OCT structural image sequences, and each group of OCT structural image sequences comprises 50 OCT structural images of a B-Scan surface;
2) and (3) screening data:
registering 50B-Scan face OCT structural images in the same group of OCT structural images by adopting a rigid registration algorithm, calculating registration accuracy by utilizing a correlation algorithm after registration, removing the whole group of OCT structural images with poor registration effect, and reserving 70 percent of screened OCT structural images, namely 100 multiplied by 30 multiplied by 0.7 group;
3) generating a training data set:
carrying out contrast imaging by using the OCTA algorithm on 100 multiplied by 30 multiplied by 0.7 groups of screened OCT structural images obtained in the step 2), wherein each group of OCT structural images obtains a B-Scan face OCTA contrast image called a label image; extracting 4B-Scan face OCT structural images from 50B-Scan face OCT structural images of a group of OCT structural images corresponding to each label image, wherein the B-Scan face OCT structural images are called input data and are matched with the label images, and the input data and the label images form a training data set required by network model training;
4) establishing a machine learning network model:
the machine learning network model adopts a deep convolutional neural network DnCNN and is composed of 20 convolutional layers, wherein the layer 1 uses 64 convolutional cores of 3 x 4 to convolve input 4 OCT structural images, 64 characteristic maps are generated, and a ReLU function is used as an activation function; the 2 nd to 19 th layers use 64 convolution cores of 3 multiplied by 64 to convolute the characteristic diagram of the previous layer, and the ReLU function is connected as output after batch normalization; the 20 th layer uses 1 64 convolution cores of 3 multiplied by 4 to convolute the characteristic diagram of the previous layer, and the output image is the output image of the network; in the network parameters, the learning rate is set to 0.001, the network parameter initialization uses the Kaiming initialization method, the number of repeated training rounds is 50, the batch size is 16, and the ratio of the training set to the test set is 7: 3;
5) training a machine learning network model:
using the machine learning network model established in the step 4), and taking input data in a training data set as input of the machine learning network model, wherein 100 × 30 × 0.7 × 0.7 groups of training sets are used for training the machine learning network model, and 100 × 30 × 0.7 × 0.3 groups of test sets are used for verifying the performance of the machine learning network model; in the training process, the training set is input into the machine learning network model in batches and repeatedly for 50 rounds of training, meanwhile, the mean square error between an output image and a label image of the machine learning network model is calculated to serve as a training error, and the training error is minimized by using an adaptive moment estimation optimization algorithm (Adam) so as to train the machine learning network model; after the training of each batch is finished, performing performance test on the machine learning network model by using the test set, and when the performance test index (mean square error between an output image and a label image) of the standby device learning network model tends to be stable, considering that the training of the machine learning network model is finished, and storing the trained machine learning network model;
6) performing OCTA imaging by using a machine learning network model:
by using the trained machine learning network model, the OCT structural image of the sample acquired by the OCTA device is used as input, the number of scanning times at the same slow axis scanning position is 4, and the output image is an OCTA contrast image, as shown in fig. 2.
Finally, it is noted that the disclosed embodiments are intended to aid in further understanding of the invention, but those skilled in the art will appreciate that: various substitutions and modifications are possible without departing from the spirit and scope of the invention and the appended claims. Therefore, the invention should not be limited to the embodiments disclosed, but the scope of the invention is defined by the appended claims.
Claims (6)
1. An optical coherent angiography imaging method based on machine learning, characterized in that the optical coherent angiography imaging method comprises the following steps:
1) generating an original data set:
acquiring an OCT three-dimensional structural image of a sample by using OCTA equipment, and generating an original data set required by network model training, wherein the original data set comprises j multiplied by k groups of OCT structural image sequences, each group of OCT structural image sequences comprises i OCT structural images with two-dimensional cross sections (B-Scan), k is the number of the sample, j is the number of scanning positions of a slow axis of each sample, i is the scanning times of the same scanning position of the slow axis of the same sample, i is a natural number larger than 4, j is a natural number larger than 50, and k is a natural number larger than 5;
2) and (3) screening data:
registering i B-Scan face OCT structural images in the same group of OCT structural images by adopting a rigid registration algorithm, calculating registration accuracy by utilizing a correlation algorithm after registration, removing the whole group of OCT structural images with poor registration effect, and reserving n groups of screened OCT structural images, wherein n is a natural number, and
3) generating a training data set:
carrying out contrast imaging by using the n groups of screened OCT structural images obtained in the step 2) by adopting an OCTA algorithm, wherein each group of OCT structural images obtains a B-Scan surface OCTA contrast image called a label image; taking m B-Scan surface OCT structural images from i B-Scan surface OCT structural images of a group of OCT structural images corresponding to each label image, wherein m is 2,3 or 4, the m is called input data, the input data is matched with the label images, and the input data and the label images form a training data set required by network model training;
4) establishing a machine learning network model:
constructing a machine learning network model, and setting hyper-parameters of the machine learning network model; and dividing the training data set into n1Group training set and n2Group test set, training set and test set independent of each other, n1And n2Are respectively natural numbers, and n1+n2＝n，
5) Training a machine learning network model:
using the machine learning network model established in the step 4), taking the input data in the training data set as the input of the machine learning network model, wherein n is used1The group training set is used for training the machine learning network model and takes n2The group test set is used for verifying the performance of the machine learning network model; in the training process, the training set is divided into a plurality of batches and repeatedly input into the machine learning network model for training for a plurality of times, the difference between the output image and the label image of the machine learning network model is judged or calculated as a training error to train the machine learning network model, after the training of each batch is finished, the test set is used for carrying out performance test on the machine learning network model, when the performance test indexes of the standby machine learning network model tend to be stable, the training of the machine learning network model is considered to be finished, and the trained machine learning network model is stored;
6) performing OCTA imaging by using a machine learning network model:
and (3) by utilizing the trained machine learning network model, taking the OCT structural image of the sample acquired by the OCTA equipment as input, wherein the output image is the OCTA contrast image.
2. The method as claimed in claim 1, wherein in step 1), when the OCTA device collects the samples, the OCT structure images of one B-Scan plane are obtained by scanning one slow axis scanning position of one sample once, the OCT structure images of the same slow axis scanning position are scanned i times, each sample has j slow axis scanning positions, there are k samples, so as to obtain i × j × k OCT structure images of the B-Scan plane, and the i × j × k OCT structure images of the B-Scan plane are divided into j × k groups of OCT structure images, each group of OCT structure images includes i OCT structure images of the B-Scan plane, that is, each slow axis scanning position corresponds to one group of OCT structure images, so as to obtain the original data set.
3. The optical coherence angiography imaging method according to claim 1, wherein in step 2), for the j × k sets of OCT structure images after registration, the registration accuracy is compared, the first n sets of OCT structure images with higher registration accuracy remain, the remaining sets are considered to have poor registration effect, and these sets of OCT structure images are rejected.
4. The optical coherence angiography imaging method according to claim 1, wherein in step 4), the machine learning network model employs a deep convolutional neural network CNN, a generative countermeasure network GAN, or a recurrent neural network RNN.
5. The method of claim 1, wherein in step 4), the hyper-parameters include the number of network layers, convolution kernel, learning rate, parameter initialization, number of training rounds, and batch size.
6. The method of claim 1, wherein in step 5), a mean square error, a structural similarity or a peak signal-to-noise ratio between the output image and the label image is used as a training error, and one of a stochastic gradient descent, an adaptive moment estimation optimization algorithm and a momentum algorithm is used to minimize the training error to train the machine learning network model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910513946.0A CN112085830B (en) | 2019-06-14 | 2019-06-14 | Optical coherence angiography imaging method based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910513946.0A CN112085830B (en) | 2019-06-14 | 2019-06-14 | Optical coherence angiography imaging method based on machine learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112085830A true CN112085830A (en) | 2020-12-15 |
| CN112085830B CN112085830B (en) | 2024-02-27 |
Family
ID=73733802
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910513946.0A Active CN112085830B (en) | 2019-06-14 | 2019-06-14 | Optical coherence angiography imaging method based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112085830B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114209278A (en) * | 2021-12-14 | 2022-03-22 | 复旦大学 | Deep learning skin disease diagnosis system based on optical coherence tomography |
| CN116313115A (en) * | 2023-05-10 | 2023-06-23 | 浙江大学 | Drug action mechanism prediction method based on mitochondrial dynamic phenotype and deep learning |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120213423A1 (en) * | 2009-05-29 | 2012-08-23 | University Of Pittsburgh -- Of The Commonwealth System Of Higher Education | Blood vessel segmentation with three dimensional spectral domain optical coherence tomography |
| CN107865642A (en) * | 2017-09-28 | 2018-04-03 | 温州医科大学 | A kind of glaucoma filtering operation avascular filtering bleb evaluation method based on OCT Angiographies |
| CN108670239A (en) * | 2018-05-22 | 2018-10-19 | 浙江大学 | A kind of the three-dimensional flow imaging method and system in feature based space |
| US20190090732A1 (en) * | 2017-09-27 | 2019-03-28 | Topcon Corporation | Ophthalmic apparatus, ophthalmic image processing method and recording medium |
-
2019
- 2019-06-14 CN CN201910513946.0A patent/CN112085830B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120213423A1 (en) * | 2009-05-29 | 2012-08-23 | University Of Pittsburgh -- Of The Commonwealth System Of Higher Education | Blood vessel segmentation with three dimensional spectral domain optical coherence tomography |
| US20190090732A1 (en) * | 2017-09-27 | 2019-03-28 | Topcon Corporation | Ophthalmic apparatus, ophthalmic image processing method and recording medium |
| CN107865642A (en) * | 2017-09-28 | 2018-04-03 | 温州医科大学 | A kind of glaucoma filtering operation avascular filtering bleb evaluation method based on OCT Angiographies |
| CN108670239A (en) * | 2018-05-22 | 2018-10-19 | 浙江大学 | A kind of the three-dimensional flow imaging method and system in feature based space |
Non-Patent Citations (1)
| Title |
|---|
| 周丽萍;李培;潘聪;郭立;丁志华;李鹏: "高灵敏、高对比度无标记三维光学微血管造影***与脑科学应用研究", 物理学报, vol. 65, no. 15 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114209278A (en) * | 2021-12-14 | 2022-03-22 | 复旦大学 | Deep learning skin disease diagnosis system based on optical coherence tomography |
| CN114209278B (en) * | 2021-12-14 | 2023-08-25 | 复旦大学 | Deep learning skin disease diagnosis system based on optical coherence tomography |
| CN116313115A (en) * | 2023-05-10 | 2023-06-23 | 浙江大学 | Drug action mechanism prediction method based on mitochondrial dynamic phenotype and deep learning |
| CN116313115B (en) * | 2023-05-10 | 2023-08-15 | 浙江大学 | Drug action mechanism prediction method based on mitochondrial dynamic phenotype and deep learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112085830B (en) | 2024-02-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10762637B2 (en) | Vascular segmentation using fully convolutional and recurrent neural networks | |
| JP5334415B2 (en) | Process, system and software for measuring mechanical strain and elastic properties of samples | |
| Johnstonbaugh et al. | A deep learning approach to photoacoustic wavefront localization in deep-tissue medium | |
| Benes et al. | Automatically designed machine vision system for the localization of CCA transverse section in ultrasound images | |
| US20230196572A1 (en) | Method and system for an end-to-end deep learning based optical coherence tomography (oct) multi retinal layer segmentation | |
| Lan et al. | Reconstruct the photoacoustic image based on deep learning with multi-frequency ring-shape transducer array | |
| EP3937755A1 (en) | Device and method for analyzing optoacoustic data, optoacoustic system and computer program | |
| Sleman et al. | A novel 3D segmentation approach for extracting retinal layers from optical coherence tomography images | |
| CN112085830B (en) | Optical coherence angiography imaging method based on machine learning | |
| Prasad et al. | DeepUCT: Complex cascaded deep learning network for improved ultrasound tomography | |
| Madhumathy et al. | Deep learning based photo acoustic imaging for non-invasive imaging | |
| Li et al. | Deep learning algorithm for generating optical coherence tomography angiography (OCTA) maps of the retinal vasculature | |
| CN114565711A (en) | Heart image reconstruction method and system based on deep learning | |
| WO2021106967A1 (en) | Ocular fundus image processing device and ocular fundus image processing program | |
| Sun et al. | An iterative gradient convolutional neural network and its application in endoscopic photoacoustic image formation from incomplete acoustic measurement | |
| Plyer et al. | Imaging the vasculature of a beating heart by dynamic speckle: the challenge of a quasiperiodic motion | |
| Kumar et al. | Improved Blood Vessels Segmentation of Retinal Image of Infants. | |
| Sultana et al. | RIMNet: image magnification network with residual block for retinal blood vessel segmentation | |
| Ma et al. | Deep‐learning segmentation method for optical coherence tomography angiography in ophthalmology | |
| Cisneros-Guzmán et al. | Segmentation of OCT and OCT-a images using convolutional neural networks | |
| Zengin et al. | Low-Resolution Retinal Image Vessel Segmentation | |
| Youssef et al. | A patch-wise deep learning approach for myocardial blood flow quantification with robustness to noise and nonrigid motion | |
| Kucukgoz et al. | Evaluation of 2D and 3D deep learning approaches for predicting visual acuity following surgery for idiopathic full-thickness macular holes in spectral domain optical coherence tomography images | |
| Gende et al. | A new generative approach for optical coherence tomography data scarcity: unpaired mutual conversion between scanning presets | |
| Kumar et al. | Design and Implementation of Anomaly Detection through Deep Residual Network |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |