CN110070510A - A kind of CNN medical image denoising method for extracting feature based on VGG-19 - Google Patents
A kind of CNN medical image denoising method for extracting feature based on VGG-19 Download PDFInfo
- Publication number
- CN110070510A CN110070510A CN201910345935.6A CN201910345935A CN110070510A CN 110070510 A CN110070510 A CN 110070510A CN 201910345935 A CN201910345935 A CN 201910345935A CN 110070510 A CN110070510 A CN 110070510A
- Authority
- CN
- China
- Prior art keywords
- vgg
- noise reduction
- image
- feature
- noise
- 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 22
- 230000009467 reduction Effects 0.000 claims abstract description 43
- 238000000605 extraction Methods 0.000 claims abstract description 31
- 238000012549 training Methods 0.000 claims abstract description 28
- 239000000284 extract Substances 0.000 claims abstract description 7
- 238000011946 reduction process Methods 0.000 claims abstract description 4
- 238000003860 storage Methods 0.000 claims description 6
- 230000000694 effects Effects 0.000 abstract description 4
- 238000013527 convolutional neural network Methods 0.000 description 15
- 230000006870 function Effects 0.000 description 13
- 238000002591 computed tomography Methods 0.000 description 10
- 238000005516 engineering process Methods 0.000 description 5
- 238000013135 deep learning Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000005855 radiation Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 239000003814 drug Substances 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 239000003795 chemical substances by application Substances 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000007850 degeneration Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000002285 radioactive effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
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
- 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
- G06T2207/10081—Computed x-ray tomography [CT]
-
- 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
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of CNN medical image denoising method that feature is extracted based on VGG-19, including building CNN noise reduction network, it extracts training image and training label image and principle is minimised as with loss function and the noise reduction network put up is trained, and noise reduction process is carried out to strong noise picture using the noise reduction network after training, export noise reduction result.The present invention is smooth in order to overcome image border after noise reduction, the problem of losing original image detail, it is proposed a kind of image using after noise reduction and image denoising method that the difference after clearly image uses VGG-19 to do feature extraction originally is optimized as loss function, the original details of image is saved to reach, improves the effect of image definition.
Description
Technical field
The present invention relates to be used as loss function after field of medical image processing more particularly to a kind of extraction feature by VGG-19
CNN medical image denoising method.
Background technique
Computed tomography (CT) imaging technique is born in nineteen seventies, is one of four big image technologies,
It since such technology has many advantages, such as that imaging time is short, high resolution, and contains much information, clinical application is very extensive.In recent years
Come, CT technology is developing rapidly and be widely used in clinical medicine domain, and also therefore people start gradually to pay attention to CT use
In the process itself there are the problem of, i.e., the excessively high problem of dose of radiation in CT scan.The x-ray tube current of CT machine and radioactive ray agent
Measure it is in a linear relationship, therefore, reduce tube current can reduce the dose of radiation that subject is subject to.But corresponding CT rebuilds figure
The quality of picture can generate serious degeneration, to affect the accuracy of clinical medicine diagnosis.Therefore how effectively to improve low
The picture quality of dosage CT becomes the hot spot of CT technical research.Currently, the method for improving low-dose CT resolution ratio is roughly divided into 3
Class: (1) string figure filters.Namely classical filtered back projection technique (FBP) (2) iterative approximation.Compare classical algorithm such as
Expectation maximization (EM) (3) image noise reduction post-processing approach.For example non-local means (NLM) algorithm, K-SVD reduce figure
As artifact, BM3D algorithm is used to image restoration.Deep learning widely and has successfully been applied in recent years regards in computer
In feel task, such as image segmentation, object detection, image super-resolution, due to data abundance, deep learning is obtained in recent years
It flourishing, nearest deep learning is also employed in medical image, such as with rolling up and neural network (CNN) image denoising,
Artifact removal.
Summary of the invention
According to set forth above in the prior art for deteriroation of image quality problem caused by reducing dose of radiation, one is proposed
Kind using the image after noise reduction and originally the difference that clearly image uses VGG-19 both to do after feature extraction is as loss function
The image denoising method optimized saves the original details of image to reach, improves the effect of image definition.
The technological means that the present invention uses is as follows:
A kind of CNN medical image denoising method for extracting feature based on VGG-19, comprising the following steps:
Step S1, CNN noise reduction network is built, and each layer deconvolution parameter is set；
Step S2, training image and training label image are extracted, principle is minimised as to the drop put up with loss function
Network of making an uproar is trained, and the loss function is the first difference for extracting feature with the second extraction feature, and first extraction is special
Sign carries out feature extraction acquisition through VGG-19 by the noise reduction result that noise reduction network exports；Described second extracts feature by training label
Image carries out feature extraction acquisition through VGG-19；
Step S3, noise reduction process is carried out to strong noise picture using the noise reduction network after training, exports noise reduction result.
Further, the first extraction feature carries out feature extraction through VGG-19 by the noise reduction result that noise reduction network exports
It obtains, characteristic extraction part of 16 convolutional layers for retaining the VGG-19 as noise reduction result is specifically included, by the VGG-
3 of 19 are complete, and articulamentum removes.
Further, the second extraction feature is by training label image to carry out feature extraction acquisition through VGG-19, specifically
It is complete by 3 of the VGG-19 including retaining characteristic extraction part of 16 convolutional layers of the VGG-19 as training label
Articulamentum removes.
The present invention also provides a kind of storage mediums comprising the program of storage, wherein described program executes above-mentioned any
Noise-reduction method described in one.
It the present invention also provides a kind of processor, is used to run program, wherein described program executes above-mentioned any one
The noise-reduction method.
Compared with the prior art, the invention has the following advantages that
The prior art all uses training label and each pixel difference for inputting picture as loss function, that is, uses
MSE mean square error is as loss function, however the mean square error for reducing each pixel merely can cause the transitions smooth at edge,
The original details of image is lost, noise reduction effect visually cannot be largely improved.The present invention is using VGG-19 as loss letter
Number, since VGG-19 network is mainly used to do 1000 kinds of image classifications, is more in line with people's vision to carry out the optimization to model
On the mode of feature is extracted from image, therefore can more highlight visual noise reduction effect, largely improve picture quality,
To overcome as simply to reduce excess smoothness problem in image border caused by pixel mean square error.
The present invention can be widely popularized in fields such as medical imaging, medical image denoisings based on the above reasons.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to do simply to introduce, it should be apparent that, the accompanying drawings in the following description is this hair
Bright some embodiments for those of ordinary skill in the art without any creative labor, can be with
It obtains other drawings based on these drawings.
Fig. 1 is noise-reduction method flow chart of the present invention.
Fig. 2 is that flow chart is extracted in present invention perception loss.
Fig. 3 a is that present invention training inputs strong noise schematic diagram.
Fig. 3 b is that present invention training inputs low noise schematic diagram.
Fig. 3 c is that the present invention generates result schematic diagram.
Specific embodiment
It should be noted that in the absence of conflict, the feature in embodiment and embodiment in the present invention can phase
Mutually combination.The present invention will be described in detail below with reference to the accompanying drawings and embodiments.
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is only
It is only a part of the embodiment of the present invention, instead of all the embodiments.It is real to the description of at least one exemplary embodiment below
It is merely illustrative on border, never as to the present invention and its application or any restrictions used.Based on the reality in the present invention
Example is applied, every other embodiment obtained by those of ordinary skill in the art without making creative efforts all belongs to
In the scope of protection of the invention.
It should be noted that term used herein above is merely to describe specific embodiment, and be not intended to restricted root
According to exemplary embodiments of the present invention.As used herein, unless the context clearly indicates otherwise, otherwise singular
Also it is intended to include plural form, additionally, it should be understood that, when in the present specification using term "comprising" and/or " packet
Include " when, indicate existing characteristics, step, operation, device, component and/or their combination.
As shown in Figure 1, the present invention provides a kind of CNN medical image denoising method for extracting feature based on VGG-19, packet
Include following steps:
Step S1, CNN noise reduction network is built, and each layer deconvolution parameter is set.CNN noise reduction model contains altogether 8 convolution
Layer, convolution kernel size is 3*3, and convolution kernel number is respectively 32,32,32,32,32,32,32,1, using noise reduction result as defeated
Out.
Step S2, training image and training label image are extracted, principle is minimised as to the drop put up with loss function
Network of making an uproar is trained, and the loss function is the first difference for extracting feature with the second extraction feature, and first extraction is special
Sign carries out feature extraction acquisition through VGG-19 by the noise reduction result that noise reduction network exports；Described second extracts feature by training label
Image carries out feature extraction acquisition through VGG-19.CNN network is trained and by gradient by backpropagation by Adam optimizer
Algorithm calculates.Specifically, VGG-19 Feature Selection Model used in the present invention is trained in advance by ImageNet, retain
Characteristic extraction part of 16 convolutional layers of VGG-19 as noise reduction result and training label, 3 of the VGG-19 are connected entirely
Layer is connect to remove.VGG-19 mono- shares 16 convolutional layers, and the convolution kernel size that convolutional layer uses is 3*3, and convolution kernel number is successively
It is 64,64,128,128,128,256,256,256,512,512,512,512,512,512,512,512；5 pond layers, pond
Changing layer is maximum pond, step-length 2*2；3 full articulamentums.The noise reduction result that the first extraction feature is exported by noise reduction network
Feature extraction acquisition is carried out through VGG-19, specifically includes feature of 16 convolutional layers for retaining the VGG-19 as noise reduction result
Part is extracted, 3 of the VGG-19 full articulamentums are removed.Described second extracts feature by training label image through VGG-19
Feature extraction acquisition is carried out, feature extraction unit of 16 convolutional layers for retaining the VGG-19 as training label is specifically included
Point, 3 of the VGG-19 full articulamentums are removed.
Specifically, the present invention realizes that optimal value, the loss function are found in gradient decline by loss function is defined as:
Wherein, x indicates that training label image, z indicate that training image, F are Frobenius norms, and CNN (x) indicates to use
CNN neural network carry out after noise reduction as a result, VGG (CNN (z)) is image zooming-out using VGG-19 network to noise reduction after complete
Feature and VGG (x) are the feature extracted using VGG-19 network to training label image, and w is the width of image, and h is image
Highly, d is the depth of image.
Step S3, noise reduction process is carried out to strong noise picture using the noise reduction network after training, exports noise reduction result.
The present invention also provides a kind of storage mediums comprising the program of storage, wherein described program executes above-mentioned any
Noise-reduction method described in one.
It the present invention also provides a kind of processor, is used to run program, wherein described program executes above-mentioned any one
The noise-reduction method.
Below by a specific application example, technical scheme is described further:
As shown in Fig. 2, using CNN noise reduction network to carry out noise reduction to image to achieve the purpose that use close to true picture
Trained VGG-19 is as feature extraction.Specific steps are as follows:
Step A: choosing 12 patients of Neusoft and scan 512 × 512 PET data of 75s as strong noise input, then into
9 convolution operations of row and activation obtain noise reduction output.Wherein preceding 8 convolution use 32 convolution kernels, use 1 for the last time
Convolution kernel.
Step B: sensing network is sent into output and obtains output feature, while corresponding 12 patients are scanned into the low of 150s
Noise PET image is sent into sensing network simultaneously and obtains label image feature, and the difference for the sensing results that the two obtains is as loss
Function trains CNN noise reduction model, i.e. loss function is L (D)=D (G (z))-D (x), and wherein L (D) indicates loss function, G (z)
Indicate the output feature that strong noise picture z is extracted by the picture after CNN noise reduction by VGG-19, strong noise picture such as Fig. 3 a
It is shown.X indicates corresponding low noise picture, and as shown in Figure 3b, D (x) is the label figure that low noise picture is extracted by VGG-19
As feature.
It repeats step A, B iteration 200000 times, after the completion of iteration, obtains trained network parameter, with the height of scanning 75s
The input of noise picture, obtained picture is obtained experimental result through the invention, as shown in Figure 3c.
Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention., rather than its limitations；To the greatest extent
Pipe present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: its according to
So be possible to modify the technical solutions described in the foregoing embodiments, or to some or all of the technical features into
Row equivalent replacement；And these are modified or replaceed, various embodiments of the present invention technology that it does not separate the essence of the corresponding technical solution
The range of scheme.
Claims (5)
1. a kind of CNN medical image denoising method for extracting feature based on VGG-19, it is characterised in that the following steps are included:
Step S1, CNN noise reduction network is built, and each layer deconvolution parameter is set；
Step S2, training image and training label image are extracted, principle is minimised as to the noise reduction network put up with loss function
It is trained, the loss function is the first difference for extracting feature with the second extraction feature, and described first extracts feature by dropping
The noise reduction result of network of making an uproar output carries out feature extraction acquisition through VGG-19；Described second extracts feature by training label image to pass through
VGG-19 carries out feature extraction acquisition；
Step S3, noise reduction process is carried out to strong noise picture using the noise reduction network after training, exports noise reduction result.
2. image denoising method according to claim 1, it is characterised in that the first extraction feature is defeated by noise reduction network
Noise reduction result out carries out feature extraction acquisition through VGG-19, specifically includes 16 convolutional layers for retaining the VGG-19 as drop
The characteristic extraction part for result of making an uproar removes 3 of the VGG-19 full articulamentums.
3. image denoising method according to claim 1 or 2, it is characterised in that described second extracts feature by training label
Image carries out feature extraction acquisition through VGG-19, specifically includes 16 convolutional layers for retaining the VGG-19 as training label
Characteristic extraction part removes 3 of the VGG-19 full articulamentums.
4. a kind of storage medium, it is characterised in that the program including storage, wherein described program is held any in claim 1-3
Noise-reduction method described in one.
5. a kind of processor, it is characterised in that for running program, wherein described program perform claim requires any one in 1-3
Noise-reduction method described in.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910345935.6A CN110070510A (en) | 2019-04-26 | 2019-04-26 | A kind of CNN medical image denoising method for extracting feature based on VGG-19 |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910345935.6A CN110070510A (en) | 2019-04-26 | 2019-04-26 | A kind of CNN medical image denoising method for extracting feature based on VGG-19 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110070510A true CN110070510A (en) | 2019-07-30 |
Family
ID=67369157
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910345935.6A Pending CN110070510A (en) | 2019-04-26 | 2019-04-26 | A kind of CNN medical image denoising method for extracting feature based on VGG-19 |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110070510A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110378438A (en) * | 2019-08-07 | 2019-10-25 | 清华大学 | Training method, device and the relevant device of Image Segmentation Model under label is fault-tolerant |
| EP4016446A1 (en) * | 2020-12-21 | 2022-06-22 | Dassault Systèmes | Intelligent denoising |
| EP4016445A1 (en) * | 2020-12-21 | 2022-06-22 | Dassault Systèmes | Detection of loss of details in a denoised image |
| WO2022225580A1 (en) | 2021-04-23 | 2022-10-27 | The Methodist Hospital | Method of ct denoising employing deep learning of multiple co-registered scans of the same person and group convolutional neural network analyses |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108364289A (en) * | 2018-03-02 | 2018-08-03 | 成都斯斐德科技有限公司 | IVOCT image vulnerable plaque automatic testing methods |
| CN108564555A (en) * | 2018-05-11 | 2018-09-21 | 中北大学 | A kind of digital image noise reduction method based on NSST and CNN |
| CN108629746A (en) * | 2018-04-24 | 2018-10-09 | 华中科技大学 | A kind of radar map image patch based on associated loss convolutional neural networks is made an uproar suppressing method |
| US20180293711A1 (en) * | 2017-04-06 | 2018-10-11 | Disney Enterprises, Inc. | Kernel-predicting convolutional neural networks for denoising |
| CN109166161A (en) * | 2018-07-04 | 2019-01-08 | 东南大学 | A kind of low-dose CT image processing system inhibiting convolutional neural networks based on noise artifacts |
-
2019
- 2019-04-26 CN CN201910345935.6A patent/CN110070510A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180293711A1 (en) * | 2017-04-06 | 2018-10-11 | Disney Enterprises, Inc. | Kernel-predicting convolutional neural networks for denoising |
| CN108364289A (en) * | 2018-03-02 | 2018-08-03 | 成都斯斐德科技有限公司 | IVOCT image vulnerable plaque automatic testing methods |
| CN108629746A (en) * | 2018-04-24 | 2018-10-09 | 华中科技大学 | A kind of radar map image patch based on associated loss convolutional neural networks is made an uproar suppressing method |
| CN108564555A (en) * | 2018-05-11 | 2018-09-21 | 中北大学 | A kind of digital image noise reduction method based on NSST and CNN |
| CN109166161A (en) * | 2018-07-04 | 2019-01-08 | 东南大学 | A kind of low-dose CT image processing system inhibiting convolutional neural networks based on noise artifacts |
Non-Patent Citations (2)
| Title |
|---|
| QINGSONG YANG ET.AL.: "Low-Dose CT Image Denoising Using a Generative Adversarial Network With Wasserstein Distance and Perceptual Loss", 《IEEE TRANSACTIONS ON MEDICAL IMAGING》 * |
| 章云港，易本顺，吴晨玥，冯雨: "基于卷积神经网络的低剂量CT图像去噪方法", 《光学学报》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110378438A (en) * | 2019-08-07 | 2019-10-25 | 清华大学 | Training method, device and the relevant device of Image Segmentation Model under label is fault-tolerant |
| EP4016446A1 (en) * | 2020-12-21 | 2022-06-22 | Dassault Systèmes | Intelligent denoising |
| EP4016445A1 (en) * | 2020-12-21 | 2022-06-22 | Dassault Systèmes | Detection of loss of details in a denoised image |
| WO2022225580A1 (en) | 2021-04-23 | 2022-10-27 | The Methodist Hospital | Method of ct denoising employing deep learning of multiple co-registered scans of the same person and group convolutional neural network analyses |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hu et al. | Artifact correction in low‐dose dental CT imaging using Wasserstein generative adversarial networks | |
| Lu et al. | Iterative reconstruction of low-dose CT based on differential sparse | |
| Xie et al. | Artifact removal using improved GoogLeNet for sparse-view CT reconstruction | |
| CN111325686B (en) | Low-dose PET three-dimensional reconstruction method based on deep learning | |
| CN107481297B (en) | CT image reconstruction method based on convolutional neural network | |
| CN110070510A (en) | A kind of CNN medical image denoising method for extracting feature based on VGG-19 | |
| CN111429379B (en) | Low-dose CT image denoising method and system based on self-supervision learning | |
| Pan et al. | Multi-domain integrative Swin transformer network for sparse-view tomographic reconstruction | |
| CN110717956B (en) | L0 norm optimization reconstruction method guided by limited angle projection superpixel | |
| Jiang et al. | Low-dose CT lung images denoising based on multiscale parallel convolution neural network | |
| Zhu et al. | Residual dense network for medical magnetic resonance images super-resolution | |
| CN109919954B (en) | Target object identification method and device | |
| CN115953494A (en) | Multi-task high-quality CT image reconstruction method based on low dose and super-resolution | |
| CN113516586A (en) | Low-dose CT image super-resolution denoising method and device | |
| WO2022246677A1 (en) | Method for reconstructing enhanced ct image | |
| CN110047054A (en) | A kind of GAN medical image denoising method for extracting feature based on VGG-19 | |
| CN106530236A (en) | Medical image processing method and system | |
| Green et al. | 3-D Neural denoising for low-dose Coronary CT Angiography (CCTA) | |
| Li et al. | An adaptive self-guided wavelet convolutional neural network with compound loss for low-dose CT denoising | |
| Chan et al. | An attention-based deep convolutional neural network for ultra-sparse-view CT reconstruction | |
| Zhu et al. | STEDNet: Swin transformer‐based encoder–decoder network for noise reduction in low‐dose CT | |
| CN114708352A (en) | Low-dose CT image recovery method and system based on unsupervised learning | |
| Yin et al. | Unpaired low-dose CT denoising via an improved cycle-consistent adversarial network with attention ensemble | |
| Pan et al. | Iterative Residual Optimization Network for Limited-angle Tomographic Reconstruction | |
| Li et al. | A prior-based metal artifact reduction algorithm for x-ray CT |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190730 |