CN108460754A - The method that low frequency ultrasound image is converted to high frequency ultrasound image - Google Patents
The method that low frequency ultrasound image is converted to high frequency ultrasound image Download PDFInfo
- Publication number
- CN108460754A CN108460754A CN201810036659.0A CN201810036659A CN108460754A CN 108460754 A CN108460754 A CN 108460754A CN 201810036659 A CN201810036659 A CN 201810036659A CN 108460754 A CN108460754 A CN 108460754A
- Authority
- CN
- China
- Prior art keywords
- image
- frequency ultrasound
- high frequency
- ultrasound image
- dimensional
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10132—Ultrasound image
- G06T2207/10136—3D ultrasound 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
- G06T2207/20021—Dividing image into blocks, subimages or windows
-
- 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
- G06T2207/30068—Mammography; Breast
-
- 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
- G06T2207/30168—Image quality inspection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Ultra Sonic Daignosis Equipment (AREA)
- Image Processing (AREA)
Abstract
The present invention relates to image processing fields, disclose a kind of method that low frequency ultrasound image is converted to high frequency ultrasound image.The method that the low frequency ultrasound image is converted to high frequency ultrasound image includes:Three-dimensional low-frequency ultrasonoscopy is obtained, including c layers of two dimensional image, n image block is chosen in the same position of each layer of two dimensional image;Wherein c, n are positive integers;The n image block includes 1 other image block of the first image block and n of same size, described first image block is a part for the layer original two dimensional image, other described image blocks are that the larger range of original two dimensional image of difference centered on described first image block is obtained through down-sampling;The neural network in the channels c*n that the c*n image block input of selection has been trained, to export high frequency ultrasound image.Low frequency ultrasound image can be converted into high frequency ultrasound image by the present invention, few to obtain noise, and depth is big, the fast image of image taking speed.
Description
Technical field
The present invention relates to image processing field, more particularly to the method that low frequency ultrasound image is converted to high frequency ultrasound image.
Background technology
Low frequency ultrasound picture noise in the prior art is big, has image taking speed fast, and penetration depth is big, and depth direction is differentiated
The low advantage of rate, but have the shortcomings that noise is big;And the advantages of high frequency ultrasound image of high frequency ultrasound system generation is noise
Few, depth direction high resolution is small but there are penetration depths, the slow disadvantage of image taking speed.
There are median filter and Optimized Bayesian Non-Local Mean for industrial quarters at present
(OBNLM) etc. simple to go ultrasonic noise method, but these methods can not improve the resolution ratio of script image, and improve
Picture quality.
Therefore, it is necessary to provide a kind of acquisition, noise is few, depth is big, the method for the fast image of image taking speed.
Invention content
The purpose of the present invention is to provide a kind of methods that low frequency ultrasound image is converted to high frequency ultrasound image, can obtain
The image that noise is few, depth is big, quality is high, image taking speed is fast.
This application provides a kind of methods that low frequency ultrasound image is converted to high frequency ultrasound image, including:
Three-dimensional low-frequency ultrasonoscopy is obtained, including c layers of two dimensional image, in the same position choosing of each layer of two dimensional image
Take n image block;Wherein c, n are positive integers;
The n image block includes the first image block and n-1 other image blocks of same size, described first image block
It is a part for the layer original two dimensional image, the n-1 other image blocks are the differences centered on described first image block
What larger range of original two dimensional image was obtained through down-sampling;
The neural network in the channels c*n that the c*n image block input of selection has been trained, to export high frequency ultrasound image.
Present invention also provides a kind of training methods of neural network, including:
Acquire the low frequency ultrasound image and high frequency ultrasound image of same position;
The low frequency ultrasound image includes c layers of two dimensional image, and n image is chosen in the same position of each layer of two dimensional image
Block;Wherein c, n are positive integers;
The n image block includes the first image block and n-1 other image blocks of same size, described first image block
It is a part for the layer original two dimensional image, the n-1 other image blocks are the differences centered on described first image block
What larger range of original two dimensional image was obtained through down-sampling;
By the neural network in the channels c*n image block input c*n of selection, the high frequency ultrasound image after output conversion;
Neural network parameter is adjusted, the high frequency ultrasound image and the high frequency ultrasound image after the conversion for making the acquisition
Average variance is minimum.
In a preference, c is odd number and c >=3.
It is described to export high frequency ultrasound image as one layer of high frequency ultrasound image in a preference.
In a preference, one layer of high frequency ultrasound image of output is corresponding with the middle layer of c layers of two dimensional image.
In a preference, described to export high frequency ultrasound image includes for high frequency ultrasound image.
In a preference, each layer of the multilayer high frequency ultrasound image of output is each with the c layers of two dimensional image respectively
Layer is corresponding.
In a preference, when being overlapped there are the edge of two image blocks, external image block carries out zero setting.
In a preference, low frequency ultrasound image and the high frequency ultrasound image of the acquisition same position are super in three-dimensional
Under sound system, the low frequency ultrasound image and high frequency ultrasound image of same position are acquired.
In a preference, the same position is the same position of patient body.
In a preference, the same position of the patient body is mammary gland position.
Embodiment of the present invention compared with prior art, at least has following difference and effect:
The present invention obtains the god for ultrasonoscopy conversion by the training to same object different frequency image contrast group
Through network parameter, to realize conversion of the low frequency ultrasound image to high frequency ultrasound image using the neural network parameter.
The conversion method is realized by GPU, each ultrasonic system collection image quality can be improved, so that ultrasonic
System formation speed is fast, and image quality is high, and depth is big, high resolution, the small image of noise.
It is appreciated that within the scope of the present invention in, above-mentioned each technical characteristic of the invention and below (such as embodiment with
Example) in specifically describe each technical characteristic between can be combined with each other, to form a new or preferred technical solution.Limit
In length, not repeated them here.
Description of the drawings
Fig. 1 is a kind of flow chart of the training method of neural network in the application first embodiment.
Fig. 2 is the stream for a kind of method that low frequency ultrasound image is converted to high frequency ultrasound image in the application second embodiment
Cheng Tu.
Specific implementation mode
In the following description, in order to make the reader understand this application better, many technical details are proposed.But this
The those of ordinary skill in field is appreciated that even if without these technical details and many variations based on the following respective embodiments
And modification, it can also realize the application technical solution claimed.
To make the object, technical solutions and advantages of the present invention clearer, the implementation below in conjunction with attached drawing to the present invention
Mode is described in further detail.
The first embodiment of the application is related to a kind of training method of neural network, and Fig. 1 is the first embodiment party of the application
A kind of flow chart of the training method of neural network in formula.As shown in Figure 1, the training method of the neural network includes following step
Suddenly:
Step 101:Under 3 D ultrasound system, the low frequency ultrasound image and high frequency ultrasound image of same position are acquired;
In one embodiment, the low frequency ultrasound image for acquiring same position includes with high frequency ultrasound image;It is super in three-dimensional
The low frequency of same position (for example 8Hz) ultrasonoscopy and high frequency (for example 15Hz) image are matched respectively under acoustic image system
Acquisition, sampling depth are subject to the most deep range of high frequency imaging.The same position includes the same position of patient body, preferably
Ground is the mammary gland position of patient;After fixing patient location, low frequency scanning is carried out to the privileged site of patient, is kept
Patient location is constant, carries out high frequency sweep to the position again.Due to the identical thus acquired patient in the position of twice sweep
The pixel of low frequency ultrasound image and the pixel of high frequency ultrasound image at same position have one-to-one relationship.
In one embodiment, the pairing ultrasonoscopy of low frequency and high frequency can be obtained using different ultrasonic body moulds,
Using the training data as neural network.
The low frequency ultrasound image includes c layers of two dimensional image, and n image is chosen in the same position of each layer of two dimensional image
Block;Wherein c, n are positive integers;
The n image block includes the first image block and n-1 other image blocks of same size, described first image block
It is a part for the layer original two dimensional image, the n-1 other image blocks are the differences centered on described first image block
What larger range of original two dimensional image was obtained through down-sampling.
In one embodiment, the same position n image block of selection in each layer of two dimensional image includes:Choose the
One image block, size X*Y;
When N >=2, centered on the first image block, the second image block is chosen successively to the n-th image block;N-th image block
Including the (n-1)th image block, each image block is contracted to X*Y.
Step 102:By the neural network in the channels c*n image block input c*n of selection, the high frequency ultrasound after output conversion
Image;
In one embodiment, the neural network includes one layer of input convolutional layer (including Relu function), several
Convolutional layer (including batch normalization and Relut function) and last layer of convolutional layer;Wherein, finally
One Ceng Juan bases will generate and original picture block (patch) image of a size.
In one embodiment, the neural network includes c*n input channels.The image block of X*Y* (c*n) is input to
The neural network in the channels c*n.It is 3*3 that first layer convolution, which uses K kernel, each kernel sizes,.Centre after first layer
The kernel of convolutional layer is 3*3, last layer of kernel size is 3*3*K, and the output of neural network is the multichannel of X*Y*c*n
Image block.
Step 103:Neural network parameter is adjusted, keeps the high frequency ultrasound image of the acquisition and the high frequency after the conversion super
The average variance of acoustic image is minimum.
The application second embodiment is related to a kind of method that low frequency ultrasound image is converted to high frequency ultrasound image, and Fig. 2 is
A kind of low frequency ultrasound image is converted to the flow chart of the method for high frequency ultrasound image in the application second embodiment.Such as Fig. 2 institutes
Show, the method that the low frequency ultrasound image is converted to high frequency ultrasound image includes:
Step 201:Three-dimensional low-frequency ultrasonoscopy is obtained, including c layers of two dimensional image, in the phase of each layer of two dimensional image
N image block is chosen with position;Wherein c, n are positive integers;
The n image block includes the first image block and n-1 other image blocks of same size, described first image block
It is a part for the layer original two dimensional image, the n-1 other image blocks are the differences centered on described first image block
What larger range of original two dimensional image was obtained through down-sampling.
In one embodiment, the same position n image block of selection in each layer of two dimensional image includes:Choose the
One image block, size X*Y;
When N >=2, centered on the first image block, the second image block is chosen successively to the n-th image block;N-th image block
Including the (n-1)th image block, each image block is contracted to X*Y.
In one embodiment, when being overlapped there are the edge of two image blocks, by outer image block zero setting.
In one embodiment, it when two dimensional image layer is located at outermost layer, repeats to the identical of the two dimensional image of its internal layer
Choose n image block in position.
Step 202:The neural network in the channels c*n that the c*n image block input of selection has been trained is super to export high frequency
Acoustic image.
In one embodiment, c is odd number and c >=3, exports one layer of high frequency ultrasound image, and the multilayer high frequency ultrasound figure
Each layer of picture is corresponding with each layer of c layers of two dimensional image respectively.
In one embodiment, c is odd number and c >=3, exports multilayer high frequency ultrasound image and the c layers of two dimensional image
Middle layer corresponds to.
In one embodiment, the neural network in the channels c*n trained passes through the god described in first embodiment
Training method through network obtains.
In one embodiment, the neural network in the channels c*n includes multiple convolutional layers
In one embodiment, the neural network in the channels the c*n input of c*n image block of selection trained, output
One tomographic image is as high frequency ultrasound image.
In one embodiment, the neural network in the channels the c*n input of c*n image block of selection trained, exports c
Tomographic image chooses middle layer as high frequency ultrasound image.
It should be noted that in the application documents of this patent, relational terms such as first and second and the like are only
For distinguishing one entity or operation from another entity or operation, without necessarily requiring or implying these entities
Or there are any actual relationship or orders between operation.Moreover, the terms "include", "comprise" or its any other
Variant is intended to non-exclusive inclusion, so that the process, method, article or equipment including a series of elements is not only
Including those elements, but also include other elements that are not explicitly listed, or further includes for this process, method, object
Product or the intrinsic element of equipment.In the absence of more restrictions, the element limited by sentence " including one ", not
There is also other identical elements in the process, method, article or apparatus that includes the element for exclusion.The application of this patent
In file, if it is mentioned that certain behavior is executed according to certain element, then refers to the meaning for executing the behavior according at least to the element, wherein
Include two kinds of situations:The behavior is executed according only to the element and the behavior is executed according to the element and other elements.
All references mentioned in the present invention is incorporated herein by reference, independent just as each document
It is incorporated as with reference to such.In addition, it should also be understood that, after reading the above teachings of the present invention, those skilled in the art can
To be made various changes or modifications to the present invention, such equivalent forms equally fall within the application range claimed.
Claims (10)
1. a kind of method that low frequency ultrasound image is converted to high frequency ultrasound image, which is characterized in that including:
Three-dimensional low-frequency ultrasonoscopy is obtained, including c layers of two dimensional image, n is chosen in the same position of each layer of two dimensional image
A image block;Wherein c, n are positive integers;
The n image block includes the first image block and n-1 other image blocks of same size, and described first image block is this
A part for layer original two dimensional image, the n-1 other image blocks are the different biggers centered on described first image block
What the original two dimensional image of range was obtained through down-sampling;
The neural network in the channels c*n that the c*n image block input of selection has been trained, to export high frequency ultrasound image.
2. a kind of training method of neural network, which is characterized in that including:
Acquire the low frequency ultrasound image and high frequency ultrasound image of same position;
The low frequency ultrasound image includes c layers of two dimensional image, and n image block is chosen in the same position of each layer of two dimensional image;
Wherein c, n are positive integers;
The n image block includes the first image block and n-1 other image blocks of same size, and described first image block is this
A part for layer original two dimensional image, the n-1 other image blocks are the different biggers centered on described first image block
What the original two dimensional image of range was obtained through down-sampling;
By the neural network in the channels c*n image block input c*n of selection, the high frequency ultrasound image after output conversion;
Neural network parameter is adjusted, the high frequency ultrasound image of the acquisition and being averaged for the high frequency ultrasound image after the conversion are made
Variance is minimum.
3. the method that low frequency ultrasound image according to claim 1 is converted to high frequency ultrasound image, which is characterized in that c is
Odd number and c >=3.
4. the method that low frequency ultrasound image according to claim 3 is converted to high frequency ultrasound image, which is characterized in that described
Output high frequency ultrasound image is one layer of high frequency ultrasound image.
5. the method that low frequency ultrasound image according to claim 4 is converted to high frequency ultrasound image, which is characterized in that output
One layer of high frequency ultrasound image it is corresponding with the middle layer of c layers of two dimensional image.
6. the method that low frequency ultrasound image according to claim 3 is converted to high frequency ultrasound image, which is characterized in that described
Output high frequency ultrasound image is multilayer high frequency ultrasound image.
7. the method that low frequency ultrasound image according to claim 6 is converted to high frequency ultrasound image, which is characterized in that output
Multilayer high frequency ultrasound image each layer it is corresponding with each layer of c layers of two dimensional image respectively.
8. the training method of neural network according to claim 2, which is characterized in that the low frequency of the acquisition same position
Ultrasonoscopy and high frequency ultrasound image are that the low frequency ultrasound image and high frequency ultrasound of same position are acquired under 3 D ultrasound system
Image.
9. the training method of neural network according to claim 2, which is characterized in that the same position is patient body
Same position.
10. the training method of neural network according to claim 9, which is characterized in that the same portion of the patient body
Position is mammary gland position.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810036659.0A CN108460754B (en) | 2018-01-15 | 2018-01-15 | Method for converting low-frequency ultrasonic image into high-frequency ultrasonic image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810036659.0A CN108460754B (en) | 2018-01-15 | 2018-01-15 | Method for converting low-frequency ultrasonic image into high-frequency ultrasonic image |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108460754A true CN108460754A (en) | 2018-08-28 |
CN108460754B CN108460754B (en) | 2020-08-18 |
Family
ID=63220219
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810036659.0A Active CN108460754B (en) | 2018-01-15 | 2018-01-15 | Method for converting low-frequency ultrasonic image into high-frequency ultrasonic image |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108460754B (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101320423A (en) * | 2008-06-26 | 2008-12-10 | 复旦大学 | Low resolution gait recognition method based on high-frequency super-resolution |
US20110118604A1 (en) * | 2009-11-19 | 2011-05-19 | Jeong Sik Kim | Providing an ultrasound spatial compound image in an ultrasound system |
CN102958445A (en) * | 2011-06-30 | 2013-03-06 | 株式会社东芝 | Ultrasonic diagnostic device, image processing device and image processing method |
CN103065299A (en) * | 2012-12-22 | 2013-04-24 | 深圳先进技术研究院 | Ultrasonic image edge extraction method and ultrasonic image edge extraction device |
CN107292815A (en) * | 2017-06-14 | 2017-10-24 | 上海联影医疗科技有限公司 | Processing method, device and the breast imaging equipment of galactophore image |
-
2018
- 2018-01-15 CN CN201810036659.0A patent/CN108460754B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101320423A (en) * | 2008-06-26 | 2008-12-10 | 复旦大学 | Low resolution gait recognition method based on high-frequency super-resolution |
US20110118604A1 (en) * | 2009-11-19 | 2011-05-19 | Jeong Sik Kim | Providing an ultrasound spatial compound image in an ultrasound system |
CN102958445A (en) * | 2011-06-30 | 2013-03-06 | 株式会社东芝 | Ultrasonic diagnostic device, image processing device and image processing method |
CN103065299A (en) * | 2012-12-22 | 2013-04-24 | 深圳先进技术研究院 | Ultrasonic image edge extraction method and ultrasonic image edge extraction device |
CN107292815A (en) * | 2017-06-14 | 2017-10-24 | 上海联影医疗科技有限公司 | Processing method, device and the breast imaging equipment of galactophore image |
Also Published As
Publication number | Publication date |
---|---|
CN108460754B (en) | 2020-08-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Li et al. | An image fusion method based on sparse representation and sum modified-Laplacian in NSCT domain | |
CN110705699B (en) | Super-resolution reconstruction method and device, electronic equipment and readable storage medium | |
CN109833061B (en) | Method for optimizing parameters of ultrasonic imaging system based on deep learning | |
Liu et al. | A medical image enhancement method using adaptive thresholding in NSCT domain combined unsharp masking | |
Sapkal et al. | Image fusion based on wavelet transform for medical application | |
JP6103395B2 (en) | Digital filter for image processing, image processing apparatus, print medium, recording medium, image processing method, and program | |
WO2014069558A1 (en) | Ultrasound diagnostic device | |
WO2021005098A1 (en) | Signal processing apparatus and method using local length scales for deblurring | |
US20190282205A1 (en) | Ultrasound imaging system and ultrasound imaging method | |
JPWO2016194177A1 (en) | Image processing apparatus, endoscope apparatus, and image processing method | |
KR20160091607A (en) | Image processing apparatus, ultrasonic apparatus including the same and control method for the same | |
Singh et al. | Redundant discrete wavelet transform based medical image fusion | |
KR100978482B1 (en) | Ultrasound system and method for forming compound image | |
Baig et al. | Compressed sensing-based distributed image compression | |
CN108460754A (en) | The method that low frequency ultrasound image is converted to high frequency ultrasound image | |
Orea-Flores et al. | Local complexity estimation based filtering method in wavelet domain for magnetic resonance imaging denoising | |
JP2021065707A (en) | Medical image processing device, learned model and medical image processing method | |
Ye et al. | Four image interpolation techniques for ultrasound breast phantom data acquired using Fischer's full field digital mammography and ultrasound system (FFDMUS): a comparative approach | |
Wencang et al. | Medical image fusion method based on wavelet multi-resolution and entropy | |
CN106859695B (en) | Q-frame T-aperture composite emission imaging method and system applied to ultrasonic probe | |
Kravchenko et al. | 3D filtering of images corrupted by mixed additive-impulsive noise | |
KR101604800B1 (en) | Method for improving quality of 2-d ultrasound image | |
CN1588445A (en) | Image fusing method based on direction filter unit | |
EP3716199B1 (en) | Signal processing apparatus and method using local length scales for deblurring | |
CN103654868A (en) | Method and system for imaging of ultrasound diagnostic device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |