CN112070767A - Micro-vessel segmentation method in microscopic image based on generating type countermeasure network - Google Patents
Micro-vessel segmentation method in microscopic image based on generating type countermeasure network Download PDFInfo
- Publication number
- CN112070767A CN112070767A CN202010947099.1A CN202010947099A CN112070767A CN 112070767 A CN112070767 A CN 112070767A CN 202010947099 A CN202010947099 A CN 202010947099A CN 112070767 A CN112070767 A CN 112070767A
- Authority
- CN
- China
- Prior art keywords
- network
- model
- microscopic image
- sample
- training
- 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
- 238000000034 method Methods 0.000 title claims abstract description 49
- 230000011218 segmentation Effects 0.000 title claims abstract description 31
- 210000004088 microvessel Anatomy 0.000 title claims description 17
- 238000012549 training Methods 0.000 claims abstract description 31
- 238000012545 processing Methods 0.000 claims abstract description 9
- 238000007781 pre-processing Methods 0.000 claims abstract description 7
- 210000001210 retinal vessel Anatomy 0.000 claims abstract description 4
- 230000008602 contraction Effects 0.000 claims description 19
- 230000009466 transformation Effects 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 7
- 230000004913 activation Effects 0.000 claims description 6
- 230000008034 disappearance Effects 0.000 claims description 6
- 239000002131 composite material Substances 0.000 claims description 3
- 230000006835 compression Effects 0.000 claims description 3
- 238000007906 compression Methods 0.000 claims description 3
- 238000012937 correction Methods 0.000 claims description 3
- 238000004880 explosion Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims 1
- 210000004204 blood vessel Anatomy 0.000 abstract description 5
- 238000003384 imaging method Methods 0.000 abstract description 2
- 239000008280 blood Substances 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000010339 dilation Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 230000002860 competitive effect Effects 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 230000002792 vascular Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- 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
- G06T5/00—Image enhancement or restoration
- G06T5/90—Dynamic range modification of images or parts thereof
-
- 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
- G06T2207/10024—Color image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10056—Microscopic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30041—Eye; Retina; Ophthalmic
-
- 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)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (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)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
A microvascular segmentation method in a microscopic image based on a generative confrontation network. The method aims to solve the problems that due to the limitation of an algorithm or low actual imaging contrast, the segmentation result of the algorithm is frequently subjected to the phenomenon of blood vessel breakage and the details of blood vessel branches are redundant. The invention comprises the following steps: establishing a training model and a sample set based on a generative confrontation network; inputting the color fundus images in the sample set into a generation model, extracting image characteristic information, and outputting a microvascular probability image under a microscopic image as a generation sample; carrying out enhancement processing on the microscopic image contrast self-adaptive histogram equalization; increasing the training data volume to perform secondary enhancement processing on the microscopic image processed by the preprocessing unit; distinguishing a real sample from a generated sample; inputting the color image of the retinal vessel to be segmented into a segmentation model, and outputting a vessel segmentation result; the invention is applied to the segmentation of the micro-blood vessels in the microscopic image.
Description
Technical Field
The invention relates to a microvascular segmentation method in a microscopic image based on a generative confrontation network.
Background
The microvascular image under the microscopic image has uneven gray scale distribution, complex vascular structure, low contrast between the target blood vessel and the image background, image noise and other influences, and the microvascular segmentation under the microscopic image faces huge challenges. Conventional segmentation methods include pattern recognition-based methods (supervised classification and unsupervised classification), matched filter-based, mathematical morphology-based, tracking-based methods, and the like.
A countermeasure Network (GAN) is generated, and the problem to be solved is how to generate new samples that fit the true sample probability distribution. The countermeasure network can be thought of as being comprised of a generative model and a discriminant model. In the training process, the two networks are alternately optimized in an iterative mode to form competitive confrontation until the two networks reach a dynamic balance. By means of counterstudy, the algorithm directly learns the distribution situation of the data. However, for complex data, such as high resolution images, it is extremely difficult to learn its pixel distribution without supervision. In 2014, Mehdi Mirza introduced constraint conditions into a generation countermeasure Network, and proposed a Conditional generation countermeasure Network (CGAN) to make a generated new sample controllable and more in line with expectations. In 2016, a Facebook AI team introduces a Deep neural Network into a generation countermeasure Network, and provides a Deep Convolution generation countermeasure Network (DCGAN), so that the training process of generating the countermeasure Network can be accelerated and the training process is more stable after the Deep Convolution generation countermeasure Network is introduced. Generation of the countermeasure network combines the neural network with the countermeasure idea, has already been applied to medical image processing, and has achieved better results in the field of medical image segmentation. Moeskops et al use GAN and dilation convolution to achieve automatic segmentation of brain MR images, and add dilation convolution instead of pooling layers to reduce loss of features in downsampling, making the segmentation result better than full convolution networks.
The method can extract most of the microvascular images, and because of the limitation of the algorithm or the low actual imaging contrast, the segmentation result of the algorithm often has the phenomenon of vessel fracture and the details of vessel branches have redundancy.
Disclosure of Invention
The invention aims to provide a microvascular segmentation method in a microscopic image based on a generative confrontation network.
The above purpose is realized by the following technical scheme:
a micro-vessel segmentation method in a microscopic image based on a generative confrontation network is characterized in that: the method comprises the following steps:
the method comprises the following steps: establishing a training model and a sample set based on a generative confrontation network;
step two: inputting the color fundus images in the sample set into a generation model, extracting image characteristic information, and outputting a microvascular probability image under a microscopic image as a generation sample;
step three: performing network training by adopting an RGB three-channel microscopic image, and performing enhancement processing on the contrast self-adaptive histogram equalization of the microscopic image;
step four: increasing the training data volume to perform secondary enhancement processing on the microscopic image processed by the preprocessing unit;
step five: simultaneously inputting the generated sample and the corresponding real sample into a discrimination model, respectively giving different labels to the real sample and the generated sample by the discrimination model, and distinguishing the real sample from the generated sample;
step six: alternately training and optimizing the generated model and the discriminant model until Nash balance between the discriminant model and the generated model is achieved, completing network training, wherein the trained training model is a segmentation model of the generated countermeasure network;
step seven: inputting the color image of the retinal vessel to be segmented into a segmentation model, and outputting a vessel segmentation result;
the generation network of the microvascular segmentation method in the microscopic image based on the generation type countermeasure network comprises an image preprocessing unit, a data enhancement unit, a generation model and a discrimination model;
the generation of the network model combines a U-type network and a resnet network;
the discrimination model adopts a deep convolution network and comprises three convolution modules, two dense connection modules and two compression layers.
The method for segmenting the microvessels in the microscopic image based on the generative confrontation network comprises a contraction path, an expansion path and an output layer.
The micro-vessel segmentation method in the microscopic image based on the generative confrontation network is characterized in that the contraction path mainly comprises a plurality of resnet volume blocks and downsampling, and a feature extraction part in the contraction path adopts the concept of a resnet network.
The method for segmenting the microvessels in the microscopic image based on the generative countermeasure network is characterized in that an expansion path mainly comprises a deconvolution block and an upsampling operation, a concatenate operation is further used before deconvolution of the deconvolution block, the image after convolution of a left side contraction path of a U-shaped network is connected to a corresponding expansion path according to channel hopping, and 2 Conv (3 x 3) + BN + ReLu combinations are adopted in convolution transformation.
According to the micro-vessel segmentation method in the microscopic image based on the generative countermeasure network, the dense connection module is composed of three BN-Relu-Conv composite layer structures.
The method for segmenting the microvessels in the microscopic image based on the generative confrontation network comprises the following specific processes of: inputting the extracted sample characteristics into two dense connection modules; in the dense connection module, the previous layer result and the current layer result are merged to be used as the input of the next layer, and if the output of the ith layer of the network is xi, the output of the ith layer of one dense connection module is expressed as:
xi=Hi(x0,x1,…xi-1)
wherein Hi([x0,x1,…xi-1]) Represents a non-linear mapping of the i-th layer, x0,x1,…xi-1The feature maps representing the outputs of the 0 … i-1 layers are merged;
and optimizing the target function by adopting a binary cross entropy loss function, and alternately optimizing and training the discriminant model and the generated model in the training process, so that the discriminant model is optimized and expressed as follows:
θDrepresenting parameters needing to be optimized of the discrimination model; l isD(D (x, g (x)),0) represents a loss for discriminating the generated sample as 0; l isD(D (x, y),1) represents a loss to discriminate a true sample as 1.
The method for segmenting the microvessels in the microscopic image based on the generative confrontation network is characterized in that a contraction path of a generative model mainly comprises a plurality of resnet volume blocks and downsampling, a feature extraction part in the contraction path adopts the concept of a resnet network, the problem of gradient disappearance or gradient explosion is solved, the integrity of information can be protected, the whole network only needs to learn a difference part between input and output, convolution transformation adopts a combination of BN + ReLu + Conv (3 x 3), batch normalization BN is used for optimization and adjustment, a linear correction unit ReLu activation function effectively reduces gradient disappearance in back propagation, and the ReLu activation function is as follows:
ReLU(x)=max(x,0)
the expansion path mainly comprises an deconvolution block and upsampling, a concatenate operation is further used before deconvolution is carried out on the deconvolution block, images after convolution of the left side contraction path of the U-shaped network are connected to the corresponding expansion path in a channel hopping mode, and 2 Conv (3 x 3) + BN + ReLu combinations are adopted in convolution transformation.
The invention has the following beneficial effects:
1. the convolution of the generation network in the anti-generation network adopts the idea of resnet, thereby greatly accelerating the training of the network and enabling the fine blood vessel branch details to be segmented.
2. The invention effectively removes the noise of the microscopic image and improves the accuracy of the blood vessel segmentation by generating the telescopic path in the network.
3. According to the method, the dense connection structure is added into the convolution block of the discrimination model, and the discrimination model has a function equivalent to a binary classifier in the countermeasure network, so that a deep convolution network with a plurality of hidden layers is constructed, abstract feature extraction is performed on an input image, an abstract expression of the abstract expression is output, namely, the probability of a real sample and a generated sample is judged, the dense connection structure is added into the convolution block in the middle of the deep convolution network on the basis of the structure, the discrimination capability of the discrimination network on the generated sample is enhanced, and the selection of the features is better guided by the countermeasure training.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention. In the drawings:
FIG. 1 is a schematic diagram of a generative countermeasure network;
FIG. 2 is a schematic diagram of a discriminant model;
FIG. 3 is a schematic structural view of a densely packed connection block;
FIG. 4 is a schematic diagram of the structure of the contraction path and the expansion path of the generative network;
Detailed Description
In a first embodiment, a method for segmenting microvessels in a microscopic image based on a generative confrontation network according to the first embodiment includes the following steps:
the method comprises the following steps: establishing a training model and a sample set based on a generative confrontation network;
step two: inputting the color fundus images in the sample set into a generation model, extracting image characteristic information, and outputting a microvascular probability image under a microscopic image as a generation sample;
step three: performing network training by adopting an RGB three-channel microscopic image, and performing enhancement processing on the contrast self-adaptive histogram equalization of the microscopic image;
step four: increasing the training data volume to perform secondary enhancement processing on the microscopic image processed by the preprocessing unit;
step five: simultaneously inputting the generated sample and the corresponding real sample into a discrimination model, respectively giving different labels to the real sample and the generated sample by the discrimination model, and distinguishing the real sample from the generated sample;
step six: alternately training and optimizing the generated model and the discriminant model until Nash balance between the discriminant model and the generated model is achieved, completing network training, wherein the trained training model is a segmentation model of the generated countermeasure network;
step seven: inputting the color image of the retinal vessel to be segmented into a segmentation model, and outputting a vessel segmentation result;
the generation network of the microvascular segmentation method in the microscopic image based on the generation type countermeasure network comprises an image preprocessing unit, a data enhancement unit, a generation model and a discrimination model;
the generation of the network model combines a U-type network and a resnet network;
the discrimination model adopts a deep convolution network and comprises three convolution modules, two dense connection modules and two compression layers.
In a second embodiment, the present embodiment is further described with respect to the method for segmenting microvessels in a microscopic image based on a generative confrontation network, where the generative network model includes a contraction path, an expansion path, and an output layer.
In a third embodiment, the present embodiment is a further description of the method for segmenting microvessels in a microscopic image based on a generative confrontation network according to the first embodiment, where the systolic path mainly includes a plurality of resnet volume blocks and downsampling, and the feature extraction part in the systolic path adopts the concept of the resnet network.
In a fourth embodiment, the present embodiment is a further description of the method for segmenting microvessels in a microscopic image based on a generative countermeasure network described in the first embodiment, where the extended path mainly consists of an deconvolution block and an upsampling, a continate operation is further used before deconvolution is performed on the deconvolution block, an image after convolution of a left side contracted path of the U-type network is linked to a corresponding extended path according to channel hopping, and convolution transformation adopts a combination of 2 Conv (3 × 3) + BN + ReLu.
In a fifth embodiment, the present embodiment is a further description of the method for segmenting microvessels in a microscopic image based on a generative countermeasure network according to the first embodiment, where the dense connection module is composed of three BN-Relu-Conv composite layer structures.
Sixth embodiment, the present embodiment is a further description of the method for segmenting microvessels in a microscopic image based on a generative confrontation network according to the first embodiment, where the specific process of the discriminant model is as follows: inputting the extracted sample characteristics into two dense connection modules; in the dense connection module, the previous layer result and the current layer result are merged to be used as the input of the next layer, and if the output of the ith layer of the network is xi, the output of the ith layer of one dense connection module is expressed as:
xi=Hi(x0,x1,…xi-1)
wherein Hi([x0,x1,…xi-1]) Represents a non-linear mapping of the i-th layer, x0,x1,…xi-1The feature maps representing the outputs of the 0 … i-1 layers are merged;
and optimizing the target function by adopting a binary cross entropy loss function, and alternately optimizing and training the discriminant model and the generated model in the training process, so that the discriminant model is optimized and expressed as follows:
θDrepresenting parameters needing to be optimized of the discrimination model; l isD(D (x, g (x)),0) represents a loss for discriminating the generated sample as 0; l isD(D (x, y),1) represents a loss to discriminate a true sample as 1.
The method for segmenting the microvessels in the microscopic image based on the generative confrontation network is characterized in that a contraction path of a generative model mainly comprises a plurality of resnet volume blocks and downsampling, a feature extraction part in the contraction path adopts the concept of a resnet network, the problem of gradient disappearance or gradient explosion is solved, the integrity of information can be protected, the whole network only needs to learn a difference part between input and output, convolution transformation adopts a combination of BN + ReLu + Conv (3 x 3), batch normalization BN is used for optimization and adjustment, a linear correction unit ReLu activation function effectively reduces gradient disappearance in back propagation, and the ReLu activation function is as follows:
ReLU(x)=max(x,0)
the expansion path mainly comprises an deconvolution block and upsampling, a concatenate operation is further used before deconvolution is carried out on the deconvolution block, images after convolution of the left side contraction path of the U-shaped network are connected to the corresponding expansion path in a channel hopping mode, and 2 Conv (3 x 3) + BN + ReLu combinations are adopted in convolution transformation.
Claims (7)
1. A micro-vessel segmentation method in a microscopic image based on a generative confrontation network is characterized in that: the method comprises the following steps:
the method comprises the following steps: establishing a training model and a sample set based on a generative confrontation network;
step two: inputting the color fundus images in the sample set into a generation model, extracting image characteristic information, and outputting a microvascular probability image under a microscopic image as a generation sample;
step three: performing network training by adopting an RGB three-channel microscopic image, and performing enhancement processing on the contrast self-adaptive histogram equalization of the microscopic image;
step four: increasing the training data volume to perform secondary enhancement processing on the microscopic image processed by the preprocessing unit;
step five: simultaneously inputting the generated sample and the corresponding real sample into a discrimination model, respectively giving different labels to the real sample and the generated sample by the discrimination model, and distinguishing the real sample from the generated sample;
step six: alternately training and optimizing the generated model and the discriminant model until Nash balance between the discriminant model and the generated model is achieved, completing network training, wherein the trained training model is a segmentation model of the generated countermeasure network;
step seven: inputting the color image of the retinal vessel to be segmented into a segmentation model, and outputting a vessel segmentation result;
the generation network of the microvascular segmentation method in the microscopic image based on the generation type countermeasure network comprises an image preprocessing unit, a data enhancement unit, a generation model and a discrimination model;
the generation of the network model combines a U-type network and a resnet network;
the discrimination model adopts a deep convolution network and comprises three convolution modules, two dense connection modules and two compression layers.
2. The method for segmenting the microvasculature in the microscopic image based on the generative confrontation network as claimed in claim 1, wherein: the generated network model comprises a contraction path, an expansion path and an output layer.
3. The method for segmenting the microvasculature in the microscopic image based on the generative confrontation network as claimed in claim 2, wherein: the contraction path mainly comprises a plurality of resnet volume blocks and downsampling, and the feature extraction part in the contraction path adopts the idea of a resnet network.
4. The method for segmenting the microvasculature in the microscopic image based on the generative confrontation network as claimed in claim 3, wherein: the expansion path mainly comprises an deconvolution block and upsampling, a continate operation is further used before deconvolution is carried out on the deconvolution block, an image after convolution of the left side contraction path of the U-shaped network is connected to the corresponding expansion path according to channel hopping, and the convolution transformation adopts a combination of 2 Conv (3 x 3) + BN + ReLu.
5. The method for segmenting the microvasculature in the microscopic image based on the generative confrontation network as claimed in claim 4, wherein: the dense connection module consists of three BN-Relu-Conv composite layer structures.
6. The method for segmenting the microvasculature in the microscopic image based on the generative confrontation network as claimed in claim 1, wherein: the specific process of the discriminant model is as follows: inputting the extracted sample characteristics into two dense connection modules; in the dense connection module, the previous layer result and the current layer result are merged to be used as the input of the next layer, and if the output of the ith layer of the network is xi, the output of the ith layer of one dense connection module is expressed as:
xi=Hi(x0,x1,…xi-1)
wherein Hi([x0,x1,…xi-1]) Represents a non-linear mapping of the i-th layer, x0,x1,…xi-1The feature maps representing the outputs of the 0 … i-1 layers are merged;
and optimizing the target function by adopting a binary cross entropy loss function, and alternately optimizing and training the discriminant model and the generated model in the training process, so that the discriminant model is optimized and expressed as follows:
θDrepresenting parameters needing to be optimized of the discrimination model; l isD(D (x, g (x)),0) represents a loss for discriminating the generated sample as 0; l isD(D (x, y),1) represents a loss to discriminate a true sample as 1.
7. The method for segmenting the microvasculature in the microscopic image based on the generative confrontation network as claimed in claim 1, wherein: the contraction path of the generation model mainly comprises a plurality of resnet volume blocks and down sampling, a feature extraction part in the contraction path adopts the idea of a resnet network, the problem of gradient disappearance or gradient explosion is solved, the integrity of information can be protected, the whole network only needs to learn the difference part of input and output, the convolution transformation adopts the combination of BN + ReLu + Conv (3 multiplied by 3), batch normalization BN is used for optimization and adjustment, a linear correction unit ReLu activation function effectively reduces gradient disappearance in back propagation, and the ReLu activation function is as follows:
ReLU(x)=max(x,0)
the expansion path mainly comprises an deconvolution block and upsampling, a concatenate operation is further used before deconvolution is carried out on the deconvolution block, images after convolution of the left side contraction path of the U-shaped network are connected to the corresponding expansion path in a channel hopping mode, and 2 Conv (3 x 3) + BN + ReLu combinations are adopted in convolution transformation.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010947099.1A CN112070767A (en) | 2020-09-10 | 2020-09-10 | Micro-vessel segmentation method in microscopic image based on generating type countermeasure network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010947099.1A CN112070767A (en) | 2020-09-10 | 2020-09-10 | Micro-vessel segmentation method in microscopic image based on generating type countermeasure network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112070767A true CN112070767A (en) | 2020-12-11 |
Family
ID=73663475
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010947099.1A Pending CN112070767A (en) | 2020-09-10 | 2020-09-10 | Micro-vessel segmentation method in microscopic image based on generating type countermeasure network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112070767A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113052931A (en) * | 2021-03-15 | 2021-06-29 | 沈阳航空航天大学 | DCE-MRI image generation method based on multi-constraint GAN |
CN113223671A (en) * | 2021-05-18 | 2021-08-06 | 浙江工业大学 | Microvascular tree generation method based on conditional generation countermeasure network and constraint rule |
CN116681627A (en) * | 2023-08-03 | 2023-09-01 | 佛山科学技术学院 | Cross-scale fusion self-adaptive underwater image generation countermeasure enhancement method |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110517250A (en) * | 2019-08-27 | 2019-11-29 | 东北大学 | It is a kind of to fight the eye ground blood vessel segmentation system for generating network based on enhancing |
CN111127447A (en) * | 2019-12-26 | 2020-05-08 | 河南工业大学 | Blood vessel segmentation network and method based on generative confrontation network |
-
2020
- 2020-09-10 CN CN202010947099.1A patent/CN112070767A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110517250A (en) * | 2019-08-27 | 2019-11-29 | 东北大学 | It is a kind of to fight the eye ground blood vessel segmentation system for generating network based on enhancing |
CN111127447A (en) * | 2019-12-26 | 2020-05-08 | 河南工业大学 | Blood vessel segmentation network and method based on generative confrontation network |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113052931A (en) * | 2021-03-15 | 2021-06-29 | 沈阳航空航天大学 | DCE-MRI image generation method based on multi-constraint GAN |
CN113223671A (en) * | 2021-05-18 | 2021-08-06 | 浙江工业大学 | Microvascular tree generation method based on conditional generation countermeasure network and constraint rule |
CN113223671B (en) * | 2021-05-18 | 2022-05-27 | 浙江工业大学 | Microvascular tree generation method based on conditional generation countermeasure network and constraint rule |
CN116681627A (en) * | 2023-08-03 | 2023-09-01 | 佛山科学技术学院 | Cross-scale fusion self-adaptive underwater image generation countermeasure enhancement method |
CN116681627B (en) * | 2023-08-03 | 2023-11-24 | 佛山科学技术学院 | Cross-scale fusion self-adaptive underwater image generation countermeasure enhancement method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112070767A (en) | Micro-vessel segmentation method in microscopic image based on generating type countermeasure network | |
CN109345538B (en) | Retinal vessel segmentation method based on convolutional neural network | |
CN110930418B (en) | Retina blood vessel segmentation method fusing W-net and conditional generation confrontation network | |
CN106683048B (en) | Image super-resolution method and device | |
CN110751636B (en) | Fundus image retinal arteriosclerosis detection method based on improved coding and decoding network | |
CN112258488A (en) | Medical image focus segmentation method | |
CN108764342B (en) | Semantic segmentation method for optic discs and optic cups in fundus image | |
CN109816612A (en) | Image enchancing method and device, computer readable storage medium | |
CN112446891A (en) | Medical image segmentation method based on U-Net network brain glioma | |
CN112085745A (en) | Retinal vessel image segmentation method of multi-channel U-shaped full convolution neural network based on balanced sampling splicing | |
CN113012163A (en) | Retina blood vessel segmentation method, equipment and storage medium based on multi-scale attention network | |
CN112288749A (en) | Skull image segmentation method based on depth iterative fusion depth learning model | |
CN113298235A (en) | Neural network architecture of multi-branch depth self-attention transformation network and implementation method | |
CN114494174A (en) | Chip welding line defect detection method and device | |
CN113763251A (en) | Image super-resolution amplification model and method thereof | |
CN115035127A (en) | Retinal vessel segmentation method based on generative confrontation network | |
CN113379606B (en) | Face super-resolution method based on pre-training generation model | |
CN114663678A (en) | ECO-GAN-based image enhancement system and method | |
CN110610480A (en) | MCASPP neural network eyeground image optic cup optic disc segmentation model based on Attention mechanism | |
CN111881803B (en) | Face recognition method based on improved YOLOv3 | |
Srinivasarao et al. | Cloud-Based LeNet-5 CNN for MRI Brain Tumor Diagnosis and Recognition. | |
CN113538363A (en) | Lung medical image segmentation method and device based on improved U-Net | |
CN111079585B (en) | Pedestrian re-identification method combining image enhancement with pseudo-twin convolutional neural network | |
CN116797618A (en) | Multi-stage segmentation method based on multi-mode MRI (magnetic resonance imaging) heart image | |
CN116779176A (en) | Method for amplifying chronic kidney disease samples after transplantation based on neural network depth characteristics |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20201211 |
|
WD01 | Invention patent application deemed withdrawn after publication |