CN114841878A - Speckle denoising method and device for optical coherence tomography image and electronic equipment - Google Patents
Speckle denoising method and device for optical coherence tomography image and electronic equipment Download PDFInfo
- Publication number
- CN114841878A CN114841878A CN202210455635.5A CN202210455635A CN114841878A CN 114841878 A CN114841878 A CN 114841878A CN 202210455635 A CN202210455635 A CN 202210455635A CN 114841878 A CN114841878 A CN 114841878A
- Authority
- CN
- China
- Prior art keywords
- image
- optical coherence
- speckle
- coherence tomography
- denoising
- 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
- 238000012014 optical coherence tomography Methods 0.000 title claims abstract description 148
- 238000000034 method Methods 0.000 title claims abstract description 55
- 238000013135 deep learning Methods 0.000 claims abstract description 69
- 238000012549 training Methods 0.000 claims abstract description 41
- 230000003287 optical effect Effects 0.000 claims abstract description 21
- 238000012545 processing Methods 0.000 claims description 42
- 238000005070 sampling Methods 0.000 claims description 10
- 230000008447 perception Effects 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 7
- 238000012935 Averaging Methods 0.000 claims description 6
- 239000000523 sample Substances 0.000 description 89
- 238000010586 diagram Methods 0.000 description 15
- 230000008569 process Effects 0.000 description 6
- 238000012360 testing method Methods 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000003384 imaging method Methods 0.000 description 3
- 238000010606 normalization Methods 0.000 description 3
- 238000011176 pooling Methods 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000010287 polarization Effects 0.000 description 2
- 238000002360 preparation method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000012472 biological sample Substances 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000003759 clinical diagnosis Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 235000013372 meat Nutrition 0.000 description 1
- 238000012634 optical imaging Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 235000013311 vegetables Nutrition 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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
- G06T2207/10072—Tomographic images
- G06T2207/10101—Optical tomography; Optical coherence tomography [OCT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The application provides an optical coherence tomography speckle denoising method, an optical coherence tomography speckle denoising device, electronic equipment and a storage medium, wherein the method comprises the following steps: obtaining an initial optical coherence layer image output by an optical coherence tomography system; inputting the initial optical coherence tomography image into a speckle denoising deep learning network model obtained by pre-training, and performing speckle denoising on the initial optical coherence tomography image by using the speckle denoising deep learning network model to obtain a target optical coherence tomography image without speckle noise. The speckle noise characteristics can be automatically extracted through the speckle noise removal deep learning network model obtained through pre-training, the speckle noise can be effectively eliminated, and the high-quality optical coherence layer image is obtained.
Description
Technical Field
The application relates to the technical field of optical imaging denoising methods, in particular to an optical coherence tomography speckle denoising method, an optical coherence tomography speckle denoising device, electronic equipment and a storage medium.
Background
Optical Coherence Tomography (OCT) can realize powerful noninvasive and high-resolution three-dimensional imaging of biological tissues and materials, and has been widely applied in the fields of biomedicine and the like. However, OCT discriminates biological tissue morphology by means of coherence of backscattered light, which inevitably generates speckle noise, resulting in a decrease in spatial resolution, and the speckle noise may mask fine but very important microstructures in the tissue, thereby negatively affecting clinical diagnosis and causing a decrease in diagnosis accuracy.
In the prior art, methods for removing speckle noise are divided into a hardware method and a software method, the hardware system in the hardware method has high design complexity, repeated scanning is also needed to reduce imaging time resolution, and images processed by the software method are too smooth to lose important detail characteristics.
Disclosure of Invention
An object of the present application is to provide a speckle denoising method and apparatus for optical coherence tomography, an electronic device, and a storage medium, which can automatically extract speckle noise characteristics, effectively eliminate speckle noise, and obtain a high-quality optical coherence layer image.
In order to achieve the above purpose, the technical solutions adopted in the embodiments of the present application are as follows:
in a first aspect, an embodiment of the present application provides a speckle denoising method for an optical coherence tomography image, where the method includes:
acquiring an initial optical coherence tomography image output by an optical coherence tomography system, wherein speckle noise exists in the initial optical coherence tomography image;
inputting the initial optical coherence tomography image into a speckle denoising deep learning network model obtained by pre-training, and performing speckle denoising on the initial optical coherence tomography image by using the speckle denoising deep learning network model to obtain a target optical coherence tomography image without speckle noise.
Optionally, the speckle denoising deep learning network model includes: an encoder and a decoder connected in sequence;
the inputting the initial optical coherence tomography image into a speckle denoising deep learning network model obtained by pre-training, and performing speckle denoising on the initial optical coherence tomography image by the speckle denoising deep learning network model to obtain a target optical coherence tomography image without speckle noise, includes:
and inputting the initial optical coherence tomography image into an encoder of the speckle denoising deep learning network model, performing downsampling processing on the initial optical coherence tomography image by the encoder, sending a downsampling processing result to a decoder of the speckle denoising deep learning network model, and performing upsampling processing on the downsampling processing result by the decoder to obtain the target optical coherence tomography image.
Optionally, before the initial optical coherence tomography image is input into a pre-trained speckle denoising deep learning network model and the speckle denoising of the initial optical coherence tomography image is performed by the speckle denoising deep learning network model, the method further includes:
constructing a sample data set, wherein the sample data set comprises a plurality of image pairs, each image pair comprises a reference image and a noise image, wherein speckle noise does not exist in the reference image, and speckle noise exists in the noise image;
constructing an initial model, the initial model comprising: the generator comprises an encoder and a decoder which are connected in sequence;
training the initial model based on the sample data set to obtain the speckle denoising deep learning network model.
Optionally, the constructing the sample data set includes:
scanning any sample object by using a speckle modulation optical coherence tomography system to obtain a plurality of initial images corresponding to the sample object, wherein a movable optical diffuse reflector is arranged between a first lens and a second lens in the speckle modulation optical coherence tomography system;
carrying out image averaging processing on the plurality of initial images to obtain a reference image corresponding to the sample object;
respectively taking each initial image as a noise image in the image pair, and respectively combining the initial image and the reference image into one image pair;
and combining the image pairs corresponding to the sample objects to obtain the sample data set.
Optionally, the training the initial model based on the sample data set to obtain the speckle denoising deep learning network model includes:
A. inputting the noise image in each image pair of the sample data set into the generator, sequentially performing down-sampling processing and up-sampling processing on the noise image by an encoder and a decoder in the generator, inputting the processed image into the discriminator, and discriminating the true probability of the processed image and the reference image by the discriminator;
B. determining first loss information of the generator according to the processed image and the reference image;
C. determining second loss information of the discriminator according to the processed image output by the discriminator and the true probability of the reference image;
D. correcting the network parameters of the generator and the network parameters of the discriminator according to the first loss information and the second loss information respectively to obtain a corrected generator and a corrected discriminator;
and A-D is executed in an iterative mode until the first loss information of the generator and the second loss information of the discriminator respectively meet preset conditions, and the generator is used as the speckle denoising deep learning network model.
Optionally, the determining first loss information of the generator according to the processed image and the reference image includes:
respectively calculating the pixel mean square loss and the perception loss of the generator according to the processed image and the reference image;
and obtaining the first loss information according to the pixel mean square loss and the perception loss.
Optionally, the determining second loss information of the discriminator according to the true probability and the false probability of the processed image and the reference image output by the discriminator includes:
and calculating second loss information of the discriminator according to the difference between the true probability of the processed image and the true probability of the reference image.
In a second aspect, an embodiment of the present application further provides an optical coherence tomography speckle denoising apparatus, where the apparatus includes:
the device comprises an acquisition module, a processing module and a processing module, wherein the acquisition module is used for acquiring an initial optical coherence tomography image output by an optical coherence tomography system, and speckle noise exists in the initial optical coherence tomography image;
and the denoising module is used for inputting the initial optical coherence tomography image into a speckle denoising deep learning network model obtained by pre-training, and performing speckle denoising on the initial optical coherence tomography image by using the speckle denoising deep learning network model to obtain a target optical coherence tomography image without speckle noise.
Optionally, the speckle denoising deep learning network model includes: an encoder and a decoder connected in sequence;
the denoising module is specifically configured to:
and inputting the initial optical coherence tomography image into an encoder of the speckle denoising deep learning network model, performing downsampling processing on the initial optical coherence tomography image by the encoder, sending a downsampling processing result to a decoder of the speckle denoising deep learning network model, and performing upsampling processing on the downsampling processing result by the decoder to obtain the target optical coherence tomography image.
Optionally, the denoising module is specifically configured to:
constructing a sample data set, wherein the sample data set comprises a plurality of image pairs, each image pair comprises a reference image and a noise image, wherein speckle noise does not exist in the reference image, and speckle noise exists in the noise image;
constructing an initial model, the initial model comprising: the generator comprises an encoder and a decoder which are connected in sequence;
training the initial model based on the sample data set to obtain the speckle denoising deep learning network model.
Optionally, the denoising module is specifically configured to:
scanning any sample object by using a speckle modulation optical coherence tomography system to obtain a plurality of initial images corresponding to the sample object, wherein a movable optical diffuse reflector is arranged between a first lens and a second lens in the speckle modulation optical coherence tomography system;
carrying out image averaging processing on the plurality of initial images to obtain a reference image corresponding to the sample object;
respectively taking each initial image as a noise image in the image pair, and respectively combining the initial image and the reference image into one image pair;
and combining the image pairs corresponding to the sample objects to obtain the sample data set.
Optionally, the denoising module is specifically configured to:
A. inputting the noise image in each image pair of the sample data set into the generator, sequentially performing down-sampling processing and up-sampling processing on the noise image by an encoder and a decoder in the generator, inputting the processed image into the discriminator, and discriminating the true probability of the processed image and the reference image by the discriminator;
B. determining first loss information of the generator according to the processed image and the reference image;
C. determining second loss information of the discriminator according to the processed image output by the discriminator and the true probability of the reference image;
D. correcting the network parameters of the generator and the network parameters of the discriminator according to the first loss information and the second loss information respectively to obtain a corrected generator and a corrected discriminator;
and A-D is executed in an iterative mode until the first loss information of the generator and the second loss information of the discriminator respectively meet preset conditions, and the generator is used as the speckle denoising deep learning network model.
Optionally, the denoising module is specifically configured to:
respectively calculating the pixel mean square loss and the perception loss of the generator according to the processed image and the reference image;
and obtaining the first loss information according to the pixel mean square loss and the perception loss.
Optionally, the denoising module is specifically configured to:
and calculating second loss information of the discriminator according to the difference between the true probability of the processed image and the true probability of the reference image.
In a third aspect, an embodiment of the present application further provides an electronic device, including: a processor, a storage medium and a bus, wherein the storage medium stores program instructions executable by the processor, when an application program runs, the processor and the storage medium communicate with each other through the bus, and the processor executes the program instructions to execute the steps of the speckle denoising method for optical coherence tomography images according to the first aspect.
In a fourth aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, the computer program being read and executing the steps of the speckle denoising method for optical coherence tomography images according to the first aspect.
The beneficial effect of this application is:
according to the speckle denoising method, device, electronic equipment and storage medium for the optical coherence tomography image, an initial optical coherence layer image output by an optical coherence tomography system is obtained; inputting the initial optical coherence tomography image into a speckle denoising deep learning network model obtained by pre-training, and performing speckle denoising on the initial optical coherence tomography image by using the speckle denoising deep learning network model to obtain a target optical coherence tomography image without speckle noise. The speckle noise characteristics can be automatically extracted through the speckle noise removal deep learning network model obtained through pre-training, the speckle noise can be effectively eliminated, and the high-quality optical coherence layer image is obtained.
Drawings
To more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and those skilled in the art can also obtain other related drawings based on the drawings without inventive efforts.
Fig. 1 is a schematic structural diagram of an application scene of a speckle denoising method for an optical coherence tomography image according to an embodiment of the present disclosure;
fig. 2 is a schematic flow chart of a speckle denoising method for an optical coherence tomography image according to an embodiment of the present disclosure;
fig. 3 is a schematic flow chart of another speckle denoising method for optical coherence tomography provided in the embodiment of the present application;
FIG. 4 is a schematic structural diagram of an initial model provided in an embodiment of the present application;
fig. 5 is a schematic flowchart of a method for constructing a sample data set according to an embodiment of the present application;
FIG. 6 is a schematic structural diagram of a speckle modulation optical coherence tomography system provided in an embodiment of the present application;
FIG. 7 is a schematic flow chart diagram illustrating a method for initial model training according to an embodiment of the present disclosure;
FIG. 8 is a system diagram illustrating initial model training provided in an embodiment of the present application;
fig. 9 is a schematic diagram of an apparatus for speckle denoising an optical coherence tomography image according to an embodiment of the present disclosure;
fig. 10 is a block diagram of an electronic device according to an embodiment of the present disclosure.
Detailed Description
In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it should be understood that the drawings in the present application are for illustrative and descriptive purposes only and are not used to limit the scope of protection of the present application. Additionally, it should be understood that the schematic drawings are not necessarily drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of the present application. It should be understood that the operations of the flow diagrams may be performed out of order, and steps without logical context may be performed in reverse order or simultaneously. One skilled in the art, under the guidance of this application, may add one or more other operations to, or remove one or more operations from, the flowchart.
In addition, the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. The components of the embodiments of the present application, generally described and illustrated in the figures herein, can be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the present application, presented in the accompanying drawings, is not intended to limit the scope of the claimed application, but is merely representative of selected embodiments of the application. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present application without making any creative effort, shall fall within the protection scope of the present application.
It should be noted that in the embodiments of the present application, the term "comprising" is used to indicate the presence of the features stated hereinafter, but does not exclude the addition of further features.
Fig. 1 is a schematic structural diagram of an application scene of the speckle denoising method for optical coherence tomography images according to the embodiment of the present disclosure, as shown in fig. 1, the present disclosure is applicable to a scene for acquiring an OCT image of any sample. The sample may be, for example, a biological sample or the like. The scene relates to an Optical Coherence Tomography (OCT) system and an electronic device. Firstly, an OCT image with speckle noise can be output by scanning any sample through the OCT system, then the OCT image with speckle noise is input into the electronic equipment, the electronic equipment carries out speckle noise removal on the OCT image by utilizing a trained speckle noise removal deep learning network model, and finally the OCT image without speckle noise is obtained.
Fig. 2 is a schematic flowchart of a speckle denoising method for an optical coherence tomography image according to an embodiment of the present disclosure, and as shown in fig. 2, the method is applied to the electronic device, which may be a terminal device with computing processing capability and display function, such as a desktop computer, a notebook computer, or the like, or may also be a server. The method comprises the following steps:
s101, acquiring an initial optical coherence layer image output by the optical coherence tomography system.
Optionally, the OCT system may include a light source, a coupler, a reference mirror, a sample arm, a reference arm, and a spectrometer, and the specific principle is as follows: the system can select a broadband light source as a light source, for example, a broadband light source with the center wavelength of 850nm and the full width at half maximum of 165nm can be adopted, light emitted by the broadband light source irradiates a sample and a reference mirror through a sample arm and a reference arm respectively through a coupler, reflected light in the two light paths is converged in the coupler under a certain condition to generate an interference signal, the interference signal is output to a spectrometer, the spectrometer transmits the detected interference signal to a computer, the computer performs gray imaging and outputs an initial OCT image, and speckle noise exists in the initial OCT image.
S102, inputting the initial optical coherence tomography image into a speckle denoising deep learning network model obtained through pre-training, and performing speckle denoising on the initial optical coherence tomography image through the speckle denoising deep learning network model to obtain a target optical coherence tomography image without speckle noise.
Optionally, the speckle denoising deep learning network model is obtained by training according to the method provided by the present application, and the trained speckle denoising deep learning network model can be combined with the OCT system in step S101, so that the initial OCT image obtained in step S101 can be input into the trained speckle denoising deep learning network model, and the initial OCT image is subjected to speckle denoising by the trained speckle denoising deep learning network model, so as to obtain a target OCT image without speckle noise.
The embodiment obtains an initial optical coherence layer image output by an optical coherence tomography system; inputting the initial optical coherence tomography image into a speckle denoising deep learning network model obtained by pre-training, and performing speckle denoising on the initial optical coherence tomography image by using the speckle denoising deep learning network model to obtain a target optical coherence tomography image without speckle noise. The speckle noise characteristics can be automatically extracted through the speckle noise removal deep learning network model obtained through pre-training, the speckle noise can be effectively eliminated, and the high-quality optical coherence layer image is obtained.
Optionally, the speckle denoising deep learning network model may include an encoder and a decoder connected in sequence.
The encoder may include a convolution layer, a pooling layer, and a dense connection block, the decoder may include an inverse convolution layer, a convolution layer, and a dense connection block, each layer may be connected by the dense connection block, an image output by the encoder may be input to the decoder for decoding, and a target image may be output by the decoder.
Optionally, in step S102, the inputting the initial optical coherence tomography image into a speckle denoising deep learning network model obtained by pre-training, and performing speckle denoising on the initial optical coherence tomography image by using the speckle denoising deep learning network model to obtain a target optical coherence tomography image without speckle noise includes:
optionally, the noisy initial OCT image is input to an encoder of the speckle denoising deep learning network model, the encoder performs downsampling on the initial optical coherence tomography image, and sends a downsampled result to a decoder of the speckle denoising deep learning network model, and the decoder performs upsampling on the downsampled result to obtain the target OCT image. Specifically, an initial OCT image is input to an encoder of a speckle denoising deep learning network model, downsampling is carried out through a convolution layer, a dense connecting block and a pooling layer in the encoder, image characteristics are extracted, a characteristic image extracted by downsampling is sent to the encoder, upsampling is carried out through the convolution layer, the dense connecting block and a reverse convolution layer in a decoder, the size of the image is recovered, the size of the image is consistent with that of an original OCT image, then the image after characteristic extraction is output and serves as a target OCT image, and the target OCT image is a noise-free OCT image subjected to speckle denoising.
The following embodiment describes the training process of the speckle denoising deep learning network model.
Fig. 3 is a schematic flow chart of another optical coherence tomography image speckle denoising method provided in this embodiment of the present application, and as shown in fig. 3, the step S102 inputs the initial optical coherence tomography image into a speckle denoising deep learning network model obtained by pre-training, and the process of training the speckle denoising deep learning network model before performing speckle denoising on the initial optical coherence tomography image by using the speckle denoising deep learning network model may include:
s201, constructing a sample data set.
Optionally, the sample data set may include a plurality of image pairs, each image pair including a reference image and a noise image, wherein speckle noise is present in the reference image and speckle noise is present in the noise image.
S202, constructing an initial model.
Optionally, the initial model may include a generator and a discriminator, wherein the generator may include an encoder and a decoder connected in sequence, and a noise-free OCT image may be generated by the encoder and the decoder in the generator.
Illustratively, the generator may be a U-Net structure consisting of an encoder and a decoder built from a dense connection network.
Alternatively, the discriminator may include convolution, an activation function, and the like, and the discriminator may discriminate a difference between an image and a true image, thereby discriminating whether the image is a true image or a false image, and outputting a true probability and a false probability.
Optionally, the generator and the discriminator may be in communication, and the image generated by the generator may be sent to the discriminator, and the discriminator may be used to discriminate whether the generated image is a true image or a false image, so that the generator may be trained based on the discrimination result of the discriminator. Fig. 4 is a schematic structural diagram of an initial model provided in an embodiment of the present application, as shown in fig. 4, the left input and output are the input and output of a generator, and the right structure for determining true and false is a schematic structural diagram of a discriminator.
S203, training the initial model based on the sample data set to obtain a speckle denoising deep learning network.
Optionally, a plurality of image pairs in the sample data set in S201 may be used, and a preset method may be used to train the initial model, specifically, a generator and a discriminator in the initial model may be trained by a preset method, so that the generator may generate a speckle-free OCT image close enough to a real image, and the discriminator may not distinguish the image generated by the generator from the real image. Meanwhile, the training discriminator can identify the difference between the image generated by the generator and the real image as much as possible, namely, the generator can generate a high-quality image, the discriminator cannot distinguish the truth of the image, and then the initial model after training is used as a speckle denoising deep learning network.
Fig. 5 is a schematic flowchart of a method for constructing a sample data set according to an embodiment of the present application, and as shown in fig. 5, the constructing a sample data set in step S201 may include:
s301, scanning any sample by adopting a speckle modulation optical coherence tomography system to obtain a plurality of initial images corresponding to the sample object.
Fig. 6 is a schematic structural diagram of a speckle modulation optical coherence tomography system provided in an embodiment of the present application, and as shown in fig. 6, the speckle modulation optical coherence tomography system includes a spectrometer, a fiber coupler, a collimator lens, a first lens, a second lens, a third lens, a fourth lens, an objective lens, glass, a two-dimensional scanning galvanometer, a first polarization controller, a second polarization controller, a first mirror surface, a second mirror surface, and a movable diffuse optical reflector.
Optionally, the speckle modulation optical coherence tomography system adds a movable optical diffuse reflector to the optical coherence tomography system, and in particular, includes a movable optical diffuse reflector between the first lens and the second lens in fig. 6, and the initial image can be acquired by moving the optical diffuse reflector while scanning the sample object, for example, by moving the optical diffuse reflector up and down.
For example, an initial image may be obtained by moving the diffuse optical reflector once, and then a plurality of initial images may be obtained by continuously moving the diffuse optical reflector a plurality of times, where the plurality of initial images obtained by scanning with the diffuse optical reflector are a plurality of images with speckle noise, and the speckle noise of the plurality of initial images are mutually incoherent, and the reference image obtained according to the processing method of step S202 may be a reference image without speckle noise according to the characteristic that the speckle noise of the plurality of initial images is mutually incoherent.
Optionally, a speckle modulation optical coherence tomography system may be respectively used to scan a plurality of sample objects to obtain a plurality of initial images corresponding to each sample, where the sample objects may include vegetables, fruits, and meat.
Illustratively, the plurality of initial images may be, for example, 100 initial images.
And S302, respectively carrying out image averaging processing on the plurality of initial images to obtain reference images corresponding to the sample objects.
Optionally, the averaging process is performed on a plurality of initial images corresponding to each sample, for example, if one sample object is scanned in the step S301 to obtain 100 initial images, the 100 images are averaged, and the average calculated image is used as the reference image corresponding to the sample object. Specifically, pixel values of a plurality of initial images may be averaged, and an image corresponding to the pixel values after the average calculation is used as a reference image corresponding to the sample object, where the reference image is a reference image without speckle noise.
S303, each of the initial images is used as a noise image in an image pair, and is combined with the reference image to form one image pair.
Optionally, for each sample object, one of the multiple initial images corresponding to the sample object is selected to be combined with the reference image corresponding to the sample object to obtain an image pair, and by using the method, the other initial images are combined with the reference image respectively to obtain multiple image pairs, where each image pair includes a noise image and a reference image.
For example, if there are 100 initial images corresponding to a sample object, there may be 100 image pairs for the sample object.
And S304, combining the image pairs corresponding to the sample objects to obtain a sample data set.
The sample object may be one sample object or a plurality of sample objects, and then, a data set of all image pair combinations corresponding to all sample objects is used as a sample data set, and sample data sets of a preset proportion may be respectively used as a training sample data set and a test sample data set, for example, 80% of the sample data sets may be used as the training sample data set, and 20% of the sample data sets may be used as the test sample data set.
For example, if there are 1000 sample objects, each sample object may obtain 100 image pairs corresponding to each sample object through the above S301 to S303, and the sample data set includes 1000 × 100 image pairs.
In the embodiment, the obtained reference image can be closer to a real image through the noisy image and the reference image obtained by the modified OCT system, so that the trained initial model is more optimized.
Fig. 7 is a schematic flowchart of a method for training an initial model provided in the embodiment of the present application, and as shown in fig. 7, the step S203 trains the initial model based on a sample data set to obtain a speckle denoising deep learning network, including:
s401, the noise image in each image pair of the sample data set is input to the generator, the encoder and decoder in the generator sequentially perform down-sampling and up-sampling on the input noise image, the processed image is input to the discriminator, and the discriminator discriminates the true probability of the processed image and the reference image.
Optionally, before the noise image in each image pair in the sample data set is input into the generator, a normalization method may be used to perform normalization processing on the training sample data set and the test sample data set in the sample data set, specifically, the noise image in each image pair and the corresponding reference image are selected to perform normalization processing, specifically, pixel values of the noise image and the reference image in each image pair may be normalized, and a grayscale image with 256 brightness levels in the image is selected through an appropriate threshold value to obtain an image that can still reflect the overall and local features of the image, and the multi-level value of the pixel is not involved in the image processing, so that the subsequent processing process of the image is simpler.
Optionally, the noise image in each image pair in the normalized training sample data set is input to the generator, the encoder in the generator performs downsampling processing on the input noise image, the decoder performs upsampling on the downsampled noise image to obtain an image after downsampling and upsampling processing, the generator inputs the image after output processing to the discriminator, the discriminator inputs a reference image corresponding to the noise image, the discriminator determines the difference between the processed image and the corresponding reference image, so as to determine whether the processed image is a true image or a false image, and outputs the true probability of the processed image and the reference image.
S402, determining first loss information of the generator according to the processed image and the reference image.
Alternatively, the first loss information of the generator may be determined by using a preset formula according to the processed image and the reference image. The first loss information may include a pixel mean square loss and a perceptual loss.
And S403, determining second loss information of the discriminator according to the true probabilities of the processed image and the reference image output by the discriminator.
Optionally, the second loss information of the discriminator may be determined by using a preset formula according to the true probabilities of the processed image and the reference image output by the discriminator.
And S404, respectively correcting the network parameters of the generator and the network parameters of the discriminator according to the first loss information and the second loss information to obtain a corrected generator and a corrected discriminator.
The network parameters of the generator may include convolution parameters, inverse convolution parameters, pooling parameters, and the like, and the network parameters of the discriminator may include convolution parameters, activation parameters, and the like.
Optionally, the steps S401 to S404 are iteratively performed, specifically, an image pair is iteratively selected from the training sample data set, the steps S401 to S404 are respectively performed, the network parameters of the generator and the discriminator are continuously corrected until the first loss information of the generator and the second loss information of the discriminator respectively satisfy preset conditions, the corrected generator and the discriminator are tested by using the test data set, and the finally corrected generator is used as the speckle denoising deep learning network model after the test is passed.
For example, the preset condition may be that the first loss information and the second loss information tend to be stable.
Fig. 8 is a schematic structural diagram of a system for initial model training provided in an embodiment of the present application, and as shown in fig. 8, the training system includes sample data set preparation and initial model training.
The specific implementation method for sample data set preparation is described in detail in the foregoing specific embodiments, and is not described herein again.
Optionally, the noise image in the training sample data set is transmitted in the forward direction and input into the initial model for training, the processed image is output, the loss between the processed image and the reference image is calculated, the network parameters of the generator and the discriminator in the initial model are optimized in the gradient descending direction through the reverse transmission loss value, and meanwhile, the trained initial model is tested by using the test sample data set.
In this embodiment, the initial model is iteratively corrected by using the sample data set, so that the corrected initial model can remove speckle noise and analyze a tiny and important microstructure masked by the speckle noise.
Optionally, the step S402 of determining the first loss information of the generator according to the processed image and the reference image includes:
optionally, the mean square loss and the perceptual loss of the pixels of the generator are respectively calculated according to the processed image and the reference image.
Alternatively, the pixel mean square loss can be calculated according to equation (one).
Wherein L is 2 Denotes the pixel mean square loss, x denotes the processed image, y denotes the reference image, G (x) i ) I-th pixel value, y, representing the processed image i And (2) the ith pixel value of the reference image is shown, n represents the total number of image pixels, and the pixel mean square loss is obtained by performing mean square error calculation on the pixel values of the processed image and the reference image according to the formula (I).
Alternatively, the perceptual loss may be calculated according to equation (two).
Wherein L is VGG Representing the perceptual loss, x representing the processed image, y representing the reference image, VGG 19 (y i ) Representing benchmarks extracted through VGG-19 network featuresIth pixel value of image, VGG 19 (G(x i ) I-th pixel value of the processed image subjected to the VGG-19 network feature extraction, w represents the width of the image, h represents the height of the image, and d represents the number of channels of the image.
Optionally, the content loss may be calculated by using a formula (three) according to the calculated mean square loss of the pixel and the perceptual loss.
L content ＝α×L 2 +β×L VGG Formula (III)
Wherein L is content Represents a content loss, L 2 Representing the mean square loss of the pixel, L VGG And a and β represent coefficients of a pixel mean square loss and a perceptual loss, and illustratively, a may be 1, and β may be 0.1, for example.
Optionally, the first loss information includes the calculated pixel mean square loss, perceptual loss, and content loss, and the network parameter of the generator is subjected to gradient optimization according to the calculated content loss.
Optionally, the step S403 determines the second loss information of the discriminator according to the true probabilities of the processed image and the reference image output by the discriminator, and includes:
alternatively, the second loss information of the discriminator is calculated according to a difference between the true probability of the processed image and the true probability of the reference image, and specifically, may be calculated according to formula (iv).
Wherein L is discriminator Denotes discriminator loss, x denotes a processed image, y denotes a reference image, and D (y) i ) Indicating the probability value D [ G (x) that the ith pixel of the reference image is judged to be true i )]Probability value representing the i-th pixel of the processed image judged to be true, 1-D [ G (x) i )]Representing the probability value of the i-th pixel of the processed image judged to be false, and calculating by using a cross entropy functionThe discriminator loss of (2) is the second loss information, and the gradient optimization is performed on the network parameters of the discriminator according to the calculated discriminator loss.
The embodiment can minimize the difference between the image generated by the generator and the real image by continuously optimizing the network parameters of the generator and the network parameters of the discriminator by using the first loss information and the second loss information, and the discriminator can maximally discriminate the true and false of the image.
Fig. 9 is a schematic diagram of an apparatus for speckle denoising an optical coherence tomography image according to an embodiment of the present disclosure, as shown in fig. 9, the apparatus includes:
an obtaining module 501, configured to obtain an initial optical coherence tomography image output by an optical coherence tomography system, where speckle noise exists in the initial optical coherence tomography image;
the denoising module 502 is configured to input the initial optical coherence tomography image into a speckle denoising deep learning network model obtained through pre-training, and perform speckle denoising on the initial optical coherence tomography image by using the speckle denoising deep learning network model to obtain a target optical coherence tomography image without speckle noise.
Optionally, the denoising module 502 is specifically configured to:
and inputting the initial optical coherence tomography image into an encoder of the speckle denoising deep learning network model, performing downsampling processing on the initial optical coherence tomography image by the encoder, sending a downsampling processing result to a decoder of the speckle denoising deep learning network model, and performing upsampling processing on the downsampling processing result by the decoder to obtain the target optical coherence tomography image.
Optionally, the denoising module 502 is specifically configured to:
constructing a sample data set, wherein the sample data set comprises a plurality of image pairs, each image pair comprises a reference image and a noise image, wherein speckle noise does not exist in the reference image, and speckle noise exists in the noise image;
constructing an initial model, the initial model comprising: the generator comprises an encoder and a decoder which are connected in sequence;
training the initial model based on the sample data set to obtain the speckle denoising deep learning network model.
Optionally, the denoising module 502 is specifically configured to:
scanning any sample object by using a speckle modulation optical coherence tomography system to obtain a plurality of initial images corresponding to the sample object, wherein a movable optical diffuse reflector is arranged between a first lens and a second lens in the speckle modulation optical coherence tomography system;
carrying out image averaging processing on the plurality of initial images to obtain a reference image corresponding to the sample object;
respectively taking each initial image as a noise image in the image pair, and respectively combining the initial image and the reference image into one image pair;
and combining the image pairs corresponding to the sample objects to obtain the sample data set.
Optionally, the denoising module 502 is specifically configured to:
A. inputting the noise image in each image pair of the sample data set into the generator, sequentially performing down-sampling processing and up-sampling processing on the noise image by an encoder and a decoder in the generator, inputting the processed image into the discriminator, and discriminating the true probability of the processed image and the reference image by the discriminator;
B. determining first loss information of the generator according to the processed image and the reference image;
C. determining second loss information of the discriminator according to the processed image output by the discriminator and the true probability of the reference image;
D. correcting the network parameters of the generator and the network parameters of the discriminator according to the first loss information and the second loss information respectively to obtain a corrected generator and a corrected discriminator;
and A-D is executed in an iterative mode until the first loss information of the generator and the second loss information of the discriminator respectively meet preset conditions, and the generator is used as the speckle denoising deep learning network model.
Optionally, the denoising module 502 is specifically configured to:
respectively calculating the pixel mean square loss and the perception loss of the generator according to the processed image and the reference image;
and obtaining the first loss information according to the pixel mean square loss and the perception loss.
Optionally, the denoising module 502 is specifically configured to:
and calculating second loss information of the discriminator according to the difference between the true probability of the processed image and the true probability of the reference image.
Fig. 10 is a block diagram of an electronic device 600 according to an embodiment of the present disclosure. As shown in fig. 10, the electronic device may include: a processor 601, a memory 602.
Optionally, a bus 603 may be further included, wherein the memory 602 is configured to store machine-readable instructions executable by the processor 601, when the electronic device 600 runs, the processor 601 and the memory 602 are in communication via the bus 603, and the machine-readable instructions are executed by the processor 601 to perform the method steps in the above method embodiments.
The present application further provides a computer-readable storage medium, on which a computer program is stored, where the computer program is executed by a processor to perform the method steps in the speckle denoising method for optical coherence tomography images.
It can be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working processes of the system and the apparatus described above may refer to corresponding processes in the method embodiments, and are not described in detail in this application. In the several embodiments provided in the present application, it should be understood that the disclosed system, apparatus and method may be implemented in other ways. The above-described apparatus embodiments are merely illustrative, and for example, the division of the modules is merely a logical division, and there may be other divisions in actual implementation, and for example, a plurality of modules 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 devices or modules through some communication interfaces, and may be in an electrical, mechanical or other form.
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 functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application or portions thereof that substantially contribute to the prior art may be embodied in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk, and various media capable of storing program codes.
The above description is only for the specific embodiments of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present application, and shall be covered by the scope of the present application.
Claims (10)
1. An optical coherence tomography speckle denoising method, comprising:
acquiring an initial optical coherence tomography image output by an optical coherence tomography system, wherein speckle noise exists in the initial optical coherence tomography image;
inputting the initial optical coherence tomography image into a speckle denoising deep learning network model obtained by pre-training, and performing speckle denoising on the initial optical coherence tomography image by using the speckle denoising deep learning network model to obtain a target optical coherence tomography image without speckle noise.
2. The speckle denoising method for optical coherence tomography images according to claim 1, wherein the speckle denoising deep learning network model comprises: an encoder and a decoder connected in sequence;
the inputting the initial optical coherence tomography image into a speckle denoising deep learning network model obtained by pre-training, and performing speckle denoising on the initial optical coherence tomography image by the speckle denoising deep learning network model to obtain a target optical coherence tomography image without speckle noise, includes:
and inputting the initial optical coherence tomography image into an encoder of the speckle denoising deep learning network model, performing downsampling processing on the initial optical coherence tomography image by the encoder, sending a downsampling processing result to a decoder of the speckle denoising deep learning network model, and performing upsampling processing on the downsampling processing result by the decoder to obtain the target optical coherence tomography image.
3. The speckle denoising method for the optical coherence tomography image according to claim 1 or 2, wherein the inputting the initial optical coherence tomography image into a pre-trained speckle denoising deep learning network model, before the speckle denoising of the initial optical coherence tomography image by the speckle denoising deep learning network model, further comprises:
constructing a sample data set, wherein the sample data set comprises a plurality of image pairs, each image pair comprises a reference image and a noise image, wherein speckle noise does not exist in the reference image, and speckle noise exists in the noise image;
constructing an initial model, the initial model comprising: the generator comprises an encoder and a decoder which are connected in sequence;
training the initial model based on the sample data set to obtain the speckle denoising deep learning network model.
4. The method for speckle denoising of optical coherence tomography images according to claim 3, wherein the constructing the sample dataset comprises:
scanning any sample object by using a speckle modulation optical coherence tomography system to obtain a plurality of initial images corresponding to the sample object, wherein a movable optical diffuse reflector is arranged between a first lens and a second lens in the speckle modulation optical coherence tomography system;
carrying out image averaging processing on the plurality of initial images to obtain a reference image corresponding to the sample object;
respectively taking each initial image as a noise image in the image pair, and respectively combining the initial image and the reference image into one image pair;
and combining the image pairs corresponding to the sample objects to obtain the sample data set.
5. The method for speckle denoising of optical coherence tomography images according to claim 3, wherein the training of the initial model based on the sample data set to obtain the speckle denoising deep learning network model comprises:
A. inputting the noise image in each image pair of the sample data set into the generator, sequentially performing down-sampling processing and up-sampling processing on the noise image by an encoder and a decoder in the generator, inputting the processed image into the discriminator, and discriminating the true probability of the processed image and the reference image by the discriminator;
B. determining first loss information of the generator according to the processed image and the reference image;
C. determining second loss information of the discriminator according to the processed image output by the discriminator and the true probability of the reference image;
D. correcting the network parameters of the generator and the network parameters of the discriminator according to the first loss information and the second loss information respectively to obtain a corrected generator and a corrected discriminator;
and A-D is executed in an iterative mode until the first loss information of the generator and the second loss information of the discriminator respectively meet preset conditions, and the generator is used as the speckle denoising deep learning network model.
6. The method for speckle denoising of optical coherence tomography images of claim 5, wherein the determining the first loss information of the generator from the processed image and the reference image comprises:
respectively calculating the pixel mean square loss and the perception loss of the generator according to the processed image and the reference image;
and obtaining the first loss information according to the pixel mean square loss and the perception loss.
7. The method for speckle denoising of optical coherence tomography images according to claim 5, wherein the determining second loss information of the discriminator according to the true probability and the false probability of the processed image and the reference image output by the discriminator comprises:
and calculating second loss information of the discriminator according to the difference between the true probability of the processed image and the true probability of the reference image.
8. An optical coherence tomography speckle denoising apparatus, comprising:
the device comprises an acquisition module, a processing module and a processing module, wherein the acquisition module is used for acquiring an initial optical coherence tomography image output by an optical coherence tomography system, and speckle noise exists in the initial optical coherence tomography image;
and the denoising module is used for inputting the initial optical coherence tomography image into a speckle denoising deep learning network model obtained by pre-training, and performing speckle denoising on the initial optical coherence tomography image by using the speckle denoising deep learning network model to obtain a target optical coherence tomography image without speckle noise.
9. An electronic device, comprising a memory and a processor, wherein the memory stores a computer program executable by the processor, and the processor implements the steps of the speckle denoising method for optical coherence tomography images according to any one of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, having stored thereon a computer program which, when being executed by a processor, carries out the steps of the method for speckle denoising optical coherence tomography images as defined in any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210455635.5A CN114841878A (en) | 2022-04-27 | 2022-04-27 | Speckle denoising method and device for optical coherence tomography image and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210455635.5A CN114841878A (en) | 2022-04-27 | 2022-04-27 | Speckle denoising method and device for optical coherence tomography image and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114841878A true CN114841878A (en) | 2022-08-02 |
Family
ID=82567429
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210455635.5A Pending CN114841878A (en) | 2022-04-27 | 2022-04-27 | Speckle denoising method and device for optical coherence tomography image and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114841878A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116385319A (en) * | 2023-05-29 | 2023-07-04 | 中国人民解放军国防科技大学 | Radar image speckle filtering method and device based on scene cognition |
| CN116754654A (en) * | 2023-08-21 | 2023-09-15 | 成都中弘轨道交通环保产业股份有限公司 | Sound barrier detection method and device based on image analysis model |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109345469A (en) * | 2018-09-07 | 2019-02-15 | 苏州大学 | It is a kind of that speckle denoising method in the OCT image of confrontation network is generated based on condition |
| CN110390647A (en) * | 2019-06-14 | 2019-10-29 | 平安科技（深圳）有限公司 | The OCT image denoising method and device for generating network are fought based on annular |
| WO2020143309A1 (en) * | 2019-01-09 | 2020-07-16 | 平安科技（深圳）有限公司 | Segmentation model training method, oct image segmentation method and apparatus, device and medium |
| US20210390668A1 (en) * | 2020-06-11 | 2021-12-16 | GE Precision Healthcare LLC | Image noise reduction method and device |
-
2022
- 2022-04-27 CN CN202210455635.5A patent/CN114841878A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109345469A (en) * | 2018-09-07 | 2019-02-15 | 苏州大学 | It is a kind of that speckle denoising method in the OCT image of confrontation network is generated based on condition |
| WO2020143309A1 (en) * | 2019-01-09 | 2020-07-16 | 平安科技（深圳）有限公司 | Segmentation model training method, oct image segmentation method and apparatus, device and medium |
| CN110390647A (en) * | 2019-06-14 | 2019-10-29 | 平安科技（深圳）有限公司 | The OCT image denoising method and device for generating network are fought based on annular |
| US20210390668A1 (en) * | 2020-06-11 | 2021-12-16 | GE Precision Healthcare LLC | Image noise reduction method and device |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116385319A (en) * | 2023-05-29 | 2023-07-04 | 中国人民解放军国防科技大学 | Radar image speckle filtering method and device based on scene cognition |
| CN116385319B (en) * | 2023-05-29 | 2023-08-15 | 中国人民解放军国防科技大学 | Radar image speckle filtering method and device based on scene cognition |
| CN116754654A (en) * | 2023-08-21 | 2023-09-15 | 成都中弘轨道交通环保产业股份有限公司 | Sound barrier detection method and device based on image analysis model |
| CN116754654B (en) * | 2023-08-21 | 2023-12-01 | 成都中弘轨道交通环保产业股份有限公司 | Sound barrier detection method and device based on image analysis model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114841878A (en) | Speckle denoising method and device for optical coherence tomography image and electronic equipment | |
| CN111062880A (en) | Underwater image real-time enhancement method based on condition generation countermeasure network | |
| US20150199796A1 (en) | Scoring and Adjusting Pixels Based on Neighborhood Relationships for Revealing Data in Images | |
| CN108042126B (en) | Improved laser speckle contrast blood flow imaging method | |
| CN111784686A (en) | Dynamic intelligent detection method, system and readable storage medium for endoscope bleeding area | |
| CN115063595A (en) | Speckle noise image processing device | |
| Vaz et al. | Re-ordering of Hadamard matrix using Fourier transform and gray-level co-occurrence matrix for compressive single-pixel imaging in low resolution images | |
| CN114494174A (en) | Chip welding line defect detection method and device | |
| CN113850796A (en) | Lung disease identification method and device based on CT data, medium and electronic equipment | |
| CN113689337A (en) | Ultrasonic image super-resolution reconstruction method and system based on generation countermeasure network | |
| CN116630154A (en) | Deconvolution super-resolution reconstruction method and device for optical coherence tomography | |
| Hu et al. | Deep learning-based inpainting of saturation artifacts in optical coherence tomography images. | |
| US20220028067A1 (en) | Systems and Methods for Quantifying Vessel Features in Ultrasound Doppler Images | |
| Mendizabal-Ruiz et al. | Probabilistic segmentation of the lumen from intravascular ultrasound radio frequency data | |
| CN115601621A (en) | Strong scattering medium active single-pixel imaging method based on condition generation countermeasure network | |
| CN116091893A (en) | Method and system for deconvolution of seismic image based on U-net network | |
| EP2693397A1 (en) | Method and apparatus for noise reduction in an imaging system | |
| CN114511615A (en) | Method and device for calibrating image | |
| CN112085657B (en) | OCT image stitching method based on binocular stereoscopic vision tracking and retinal vascular characteristics | |
| CN114663424A (en) | Endoscope video auxiliary diagnosis method, system, equipment and medium based on edge cloud cooperation | |
| CN112949585A (en) | Identification method and device for blood vessels of fundus image, electronic equipment and storage medium | |
| KR101024857B1 (en) | Ultrasound system and method for performing color modeling processing on three-dimensional ultrasound image | |
| CN116797457B (en) | Method and system for simultaneously realizing super-resolution and artifact removal of magnetic resonance image | |
| CN117495687B (en) | Underwater image enhancement method | |
| CN117541499A (en) | Image enhancement method based on photoacoustic microscopy imaging |
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 |