CN110097511B - Image noise reduction method - Google Patents
Image noise reduction method Download PDFInfo
- Publication number
- CN110097511B CN110097511B CN201910290332.0A CN201910290332A CN110097511B CN 110097511 B CN110097511 B CN 110097511B CN 201910290332 A CN201910290332 A CN 201910290332A CN 110097511 B CN110097511 B CN 110097511B
- Authority
- CN
- China
- Prior art keywords
- noise reduction
- image
- data block
- noise
- point
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 230000002093 peripheral effect Effects 0.000 claims description 10
- 239000000126 substance Substances 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000007429 general method Methods 0.000 description 1
- 238000011946 reduction process Methods 0.000 description 1
Images
Classifications
-
- G06T5/70—
Landscapes
- Image Processing (AREA)
- Facsimile Image Signal Circuits (AREA)
Abstract
The invention discloses an image noise reduction method, which comprises the following steps: s01: selecting A data blocks taking the data block where the noise reduction point is located as the center in the image to be denoised; s02: obtaining a texture intensity value of a noise reduction point; s03: determining a noise reduction parameter h of a noise reduction point according to the texture intensity value; s04: respectively determining the noise reduction weights W (k) of A-1 surrounding data blocks according to the noise reduction parameter h; k represents the kth surrounding data block; s05: calculating the pixel value of the noise-reduced point after noise reduction according to the noise reduction weight and outputting the pixel value; s06: and sequentially selecting noise reduction points in a row-by-row and column-by-column selection mode, repeating the steps S01-S05, reducing noise of all pixel units in the noise-reduced image, and outputting the noise-reduced image. The image noise reduction method provided by the invention can effectively reduce the image noise and improve the image quality of the image on the basis of keeping the image details.
Description
Technical Field
The invention relates to the field of image processing, in particular to an image denoising method.
Background
With the development of society, CCD/CMOS image sensors have recently gained wide attention and application. Image sensors generally adopt a certain mode to collect image data, and a BGR mode and a CFA mode are commonly used. The BGR mode is an image data mode capable of directly performing processing such as display, compression and the like, and a pixel point is determined by three primary Color values of R (red), G (green) and B (blue) together.
For images in Bayer format or other formats, common noise reduction algorithms include spatial noise reduction and transform domain noise reduction. The common algorithms for spatial domain noise reduction include median filtering, mean filtering, gaussian template noise reduction algorithm, etc., and the transform domain noise reduction is to transform the image into another domain, such as a frequency domain, and transform the image back after performing noise reduction processing. In any algorithm, most of the algorithms are processed by color channels respectively, and correlation among the channels is ignored, so that the noise reduction effect is limited, and meanwhile, the loss of image edge details is serious.
Disclosure of Invention
The invention aims to provide an image noise reduction method which can effectively reduce image noise and improve the image quality of an image on the basis of keeping image details.
In order to achieve the purpose, the invention adopts the following technical scheme: an image noise reduction method comprising the steps of:
s01: selecting A data blocks with the data block where the noise reduction point is located as the center in the image to be denoised, wherein the data blocks comprise B multiplied by B pixel units, the data block containing the noise reduction point is the current data block, and the rest A-1 data blocks are peripheral data blocks; a and B are integers, and 1 is more than B; a is more than or equal to 9 and less than or equal to 121;
s02: obtaining a texture intensity value of a noise reduction point;
s03: determining a noise reduction parameter h of a noise reduction point according to the texture intensity value;
s04: respectively determining the noise reduction weights W (k) of A-1 surrounding data blocks according to the noise reduction parameter h; k represents the kth surrounding data block;
s05: calculating the pixel value of the noise-reduced point after noise reduction according to the noise reduction weight and outputting the pixel value;
s06: and sequentially selecting noise reduction points in a row-by-row and column-by-column selection mode, repeating the steps S01-S05, reducing noise of all pixel units in the image to be denoised, and outputting the denoised image after noise reduction.
Further, the specific method for obtaining the texture intensity value of the noise reduction point in step S02 is as follows:
s021: calculating the similarity distance between the A-1 surrounding data blocks and the current data block;
s022: and obtaining the texture intensity value of the noise reduction point according to the maximum similarity distance.
Further, the similarity distance between the kth surrounding data block and the current data block in the step S021
Wherein, Px(k) Representing a pixel value in an xth pixel unit in a kth surrounding data block; pcxRepresenting the pixel value in the x-th pixel unit in the current data block; k and x are integers, and x is more than or equal to 1 and less than or equal to B; k is more than or equal to 1 and less than or equal to A.
Further, the texture intensity value Td (i, j) ═ a × maxdist (k) + (1-a) × PreTd for noise reduction in the step S022; wherein a ∈ [0,1], PreTd ═ max (Td (i-1, j-1), Td (i-1, j), Td (i, j-1), Td (i, j + 1)); i represents that the noise reduction point is positioned on the ith row of the image, and j represents that the noise reduction point is positioned on the jth column of the image, wherein i and j are positive integers larger than 1.
Further, in step S04, the noise reduction weight w (k) is calculated by interpolation.
Further, the specific method for calculating the weight w (k) by using the difference method in step S04 is as follows: noise reduction weight of kth surrounding data block
Wherein the content of the first and second substances,
b represents the kth surrounding data block and the current numberBased on the similarity distance of the blocks, Hlow and Hhigh respectively represent the minimum value and the maximum value of the set noise reduction parameter h.
Further, the pixel value after noise reduction in the step S05 is
P (k) represents the pixel value of the pixel unit at the position corresponding to the noise reduction point in the kth peripheral data block.
Further, the data block includes 2 × 2 pixel units, and the noise reduction point is a pixel unit at a lower right corner of the current data block.
Further, said a is equal to 15.
Further, the image to be denoised is a bayer image.
The invention has the beneficial effects that: the noise reduction method can be carried out before the demosaicing of the Bayer image, and simultaneously reduces the brightness noise and the chrominance noise of the Bayer image; the invention utilizes the similarity between the data block where the noise reduction point is located and the surrounding data blocks to reduce noise, and considers the correlation among color channels; the invention adopts a weighting method to well maintain the detail information of the image; the invention adopts different noise reduction parameters for different texture intensity areas, thereby further protecting the details of the image.
Drawings
FIG. 1 is a flow chart of an image denoising method according to the present invention.
Fig. 2 is a schematic diagram of a data block in embodiment 1.
FIG. 3 is a schematic diagram of the selected location of PreTd in example 1.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, embodiments of the present invention are described in detail below with reference to the accompanying drawings.
As shown in fig. 1, the image denoising method provided by the present invention includes the following steps:
s01: selecting A data blocks which take a data block where a noise reduction point is located as a center from an image to be denoised, wherein the data blocks comprise B multiplied by B pixel units, the data block containing the noise reduction point is a current data block, and the rest A-1 data blocks are peripheral data blocks; a and B are integers, and 1 is more than B; a is more than or equal to 9 and less than or equal to 121. When A is 9, the current data block and the surrounding data blocks form a 3 × 3 data block array, and the current data block is located at the center of the array; when a is 121, the current data block and the surrounding data blocks form an 11 × 11 data block array, and the current data block is located at the center position of the array. Specifically, the denoising method can be used for denoising a Bayer image, each data block in the Bayer image comprises two green pixel units, one red pixel unit and one blue pixel unit, and the denoising point in the method can be set as a point at a fixed position in the data block as the denoising point because the denoising point moves in a row-by-column manner.
S02: and acquiring the texture intensity value of the noise reduction point. The invention utilizes the similarity of the noise reduction point and the pixel units of the same color channel around to reduce noise, and the similarity is calculated by calculating the similarity of the data block where the noise reduction point is located and the surrounding data block, and can be represented by the similarity distance between the surrounding data block and the current data block:
dist (k) denotes the similarity distance, P, between the kth surrounding block and the current blockx(k) Representing a pixel value in an xth pixel unit in a kth surrounding data block; pcxRepresenting the pixel value in the x-th pixel unit in the current data block; k and x are integers, and x is more than or equal to 1 and less than or equal to B; k is more than or equal to 1 and less than or equal to A. After the similarity distance between each surrounding data block and the current data block is calculated, the maximum similarity distance is found, which represents the texture intensity of the area where the a data blocks are located to a certain extent, and in order to make the texture intensity more natural, the texture intensity value Td of the noise reduction point can be further obtained.
Specifically, the texture intensity value Td (i, j) of the noise reduction point is a × maxdist (k) + (1-a) × PreTd; wherein a ∈ [0,1], PreTd ═ max (Td (i-1, j-1), Td (i-1, j), Td (i, j-1), Td (i, j + 1)); i represents that the noise reduction point is positioned on the ith row of the image, and j represents that the noise reduction point is positioned on the jth column of the image, wherein i and j are positive integers larger than 1.
S03: and determining a noise reduction parameter h of the noise reduction point according to the texture intensity value. The noise reduction parameter controls the noise reduction intensity, in the image denoising process, a fixed noise reduction parameter is usually used in an image, in the invention, a noise reduction point can have different noise reduction parameters according to the current texture intensity value, the smaller the texture intensity value Td is, the more likely the noise reduction point is in a smooth region, the larger the noise reduction parameter needs to be selected, and vice versa. Meanwhile, in order to prevent the noise reduction degree difference of each region of the whole image from being too large, the minimum value Hlow and the maximum value Hhigh of the noise reduction parameter are preset, and the actually selected noise reduction parameter h of each noise reduction point belongs to Hlow and Hhigh.
S04: respectively determining the noise reduction weights W (k) of A-1 surrounding data blocks according to the noise reduction parameter h; k denotes the kth surrounding data block. Under the condition that a noise reduction parameter h is determined, the denoising weight of each surrounding data block is calculated by adopting a Gaussian function:
it can be seen that the noise reduction parameter h controls the smoothness of the noise-reduced image, and the smaller h is, the smoother is.
Specifically, the noise reduction weight of each peripheral data block may be calculated by a difference method: assuming that the image is 8-bit image data, and the noise reduction parameter h is determined, the weight lookup table at Hlow and Hhigh is calculated as follows:
For h value between Hlow and Hhigh, the weight value is calculated by the following interpolation method,
S05: calculating the pixel value after noise reduction of the noise reduction point according to the noise reduction weight and outputting:
p (k) represents the pixel value of the pixel unit at the position corresponding to the noise reduction point in the kth peripheral data block.
S06: and sequentially selecting noise reduction points in a row-by-row and column-by-column selection mode, repeating the steps S01-S05, performing noise reduction on all pixel units in the image, and outputting the image after noise reduction.
The image denoising method provided by the invention effectively reduces the noise of the image and improves the image quality on the basis of keeping the image details. The invention is further illustrated by the following specific examples:
example 1
Referring to fig. 1-3, an image denoising method according to the present invention, specifically for denoising a bayer image, includes the following steps:
s01: in the bayer image, 15 data blocks centered on the data block where the noise reduction point is located are selected, as shown in fig. 2, where the data blocks include 2 × 2 pixel units, the data block Pc containing the noise reduction point is the current data block, the remaining 14 data blocks are peripheral data blocks, and the pixel unit in the lower right corner of the data block may be set as the noise reduction point, as shown in the position of 4 in fig. 2. In the invention, the noise reduction process needs to move the noise reduction points row by row, and therefore, each pixel unit in the image becomes the noise reduction point in the moving process.
S02: and acquiring the texture intensity value of the noise reduction point. The specific method comprises the following steps:
s021: and calculating the similarity distance between the A-1 surrounding data blocks and the current data block. The similarity degree between the current data block and the surrounding data blocks, i.e. the similarity distance Dist, is calculated according to the following formula (1), but in the implementation, the algorithm has square and square, and the calculation amount is large, so the Dist is calculated by using the following formula (2):
wherein k ∈ [1,15 ]]，Px(k) Representing a pixel value in an xth pixel unit in a kth surrounding data block; pcxRepresenting the pixel value in the xth pixel unit in the current data block. x is an integer, and x is more than or equal to 1 and less than or equal to 4.
S022: and obtaining the texture intensity value of the noise reduction point according to the maximum similarity distance. Firstly, finding out the maximum value in the 15 similarity distances, namely MaxdIst, wherein the value represents the texture intensity of the area to a certain extent, and in order to enable the texture intensity to change more naturally, combining with a figure 3, obtaining the current texture intensity value by adopting the following formula;
td (i, j) ═ a × maxdist (k) + (1-a) × PreTd; where a ∈ [0,1], PreTd ═ max (Td (i-1, j-1), Td (i-1, j), Td (i, j-1), Td (i, j +1)), i indicates that the noise reduction point is located in the ith row of the bayer image, j indicates that the noise reduction point is located in the jth column of the bayer image, and a is a preset value. The initial texture intensity value is set according to an empirical value, and after the values of Td (I-1, j-1), Td (I-1, j), Td (I, j-1) and Td (I, j +1) are determined, the texture intensity value for noise reduction is calculated according to the formula, wherein the values of Td (I-1, j-1), Td (I-1, j), Td (I, j-1) and Td (I, j +1) are also the texture intensity values obtained by performing noise reduction on the pixel unit of the j-1 th row and the j-1 th column of the I-1 th row, the pixel unit of the j-1 th row and the j-1 th column of the I-th row and the pixel unit of the j +1 th row respectively according to the formula.
S03: and determining a noise reduction parameter h of the noise reduction point according to the texture intensity value. The noise reduction parameter h controls the noise reduction intensity, and the general method is that one h value is used for the whole Bayer image, in the invention, each noise reduction point can have different noise reduction intensities according to the current texture intensity value, the smaller the texture intensity value Td of the noise reduction point is, the more likely the noise reduction point is in a smooth region, the larger the noise reduction intensity is selected, namely the larger the noise reduction parameter h is; and vice versa. Meanwhile, in order to prevent the noise reduction degree difference of each region of the whole image from being too large, the maximum and minimum values, namely Hlow and Hhigh, are preset, and the noise reduction parameter h ∈ [ Hlow and Hhigh ] determined according to the texture intensity value is determined by the method.
S04: determining the noise reduction weights W (k) of 14 surrounding data blocks according to the noise reduction parameter h; k denotes the kth surrounding data block. Specifically, the noise reduction weight of each peripheral data block may be calculated by a difference method: assuming that the bayer image is 8-bit image data, when the noise reduction parameter h is determined, the weight lookup table at Hlow and Hhigh is calculated as follows:
For h value between Hlow and Hhigh, the weight value is calculated by the following interpolation method,
S05: and calculating and outputting the pixel value after noise reduction of the noise reduction point according to the noise reduction weight. The pixel value after noise reduction is
P (k) represents the pixel value of the pixel unit at the position corresponding to the noise reduction point in the kth peripheral data block.
S06: the noise reduction points are sequentially selected in a row-by-row and column-by-column selection manner, and the above steps S01 to S05 are repeated to reduce noise of all pixel cells in the bayer image, and the bayer image after the noise reduction is output.
The noise reduction method is carried out before the demosaicing of the Bayer image, and can simultaneously reduce the brightness noise and the chrominance noise of the Bayer image; the invention utilizes the similarity between the data block where the noise reduction point is located and the surrounding data blocks to reduce noise, and considers the correlation among color channels; the invention adopts a weighting method to well keep the detail information of the Bayer image; the invention adopts different noise reduction parameters for different texture intensity areas, thereby further protecting the details of the Bayer image.
The above description is only a preferred embodiment of the present invention, and the embodiment is not intended to limit the scope of the present invention, so that all equivalent structural changes made by using the contents of the specification and the drawings of the present invention should be included in the scope of the appended claims.
Claims (9)
1. An image noise reduction method, comprising the steps of:
s01: selecting A data blocks with the data block where the noise reduction point is located as the center in the image to be denoised, wherein the data blocks comprise B multiplied by B pixel units, the data block containing the noise reduction point is the current data block, and the rest A-1 data blocks are peripheral data blocks; a and B are integers, and 1 is more than B; a is more than or equal to 9 and less than or equal to 121;
s02: obtaining a texture intensity value of a noise reduction point;
s03: determining a noise reduction parameter h of a noise reduction point according to the texture intensity value; wherein h belongs to [ Hlow, Hhigh ], and Hlow and Hhigh respectively represent the minimum value and the maximum value of the set noise reduction parameter h;
s04: respectively determining the noise reduction weights W (k) of A-1 surrounding data blocks according to the noise reduction parameter h; k represents the kth surrounding data block; noise reduction weight of kth surrounding data block
Wherein the content of the first and second substances,
b represents the similarity distance between the kth surrounding data block and the current data block, and Hlow and Hhigh respectively represent the minimum value and the maximum value of the set noise reduction parameter h;
s05: calculating the pixel value of the noise-reduced point after noise reduction according to the noise reduction weight and outputting the pixel value;
s06: and sequentially selecting noise reduction points in a row-by-row and column-by-column selection mode, repeating the steps S01-S05, reducing noise of all pixel units in the image to be denoised, and outputting the denoised image.
2. The method for reducing image noise according to claim 1, wherein the specific method for obtaining the texture intensity value for reducing noise in step S02 is as follows:
s021: calculating the similarity distance between the A-1 surrounding data blocks and the current data block;
s022: and obtaining the texture intensity value of the noise reduction point according to the maximum similarity distance.
3. The method as claimed in claim 2, wherein the similarity distance between the kth surrounding data block and the current data block in step S021
Wherein, Px(k) Representing a pixel value in an xth pixel unit in a kth surrounding data block; pcxRepresenting the pixel value in the x-th pixel unit in the current data block; k and x are integers, and x is more than or equal to 1 and less than or equal to B; k is more than or equal to 1 and less than or equal to A.
4. An image denoising method according to claim 3, wherein the texture intensity value Td (i, j) ═ a × maxdist (k) ++ (1-a) × PreTd of the noise reduction point in step S022; wherein a ∈ [0,1], PreTd ═ max (Td (i-1, j-1), Td (i-1, j), Td (i, j-1), Td (i, j + 1)); i represents that the noise reduction point is positioned on the ith row of the image, and j represents that the noise reduction point is positioned on the jth column of the image, wherein i and j are positive integers larger than 1.
5. An image denoising method according to claim 2, wherein the denoising weight w (k) is calculated in step S04 by interpolation.
6. An image denoising method according to claim 1, wherein the pixel value after denoising in step S05 is
P (k) represents the pixel value of the pixel unit at the position corresponding to the noise reduction point in the kth peripheral data block.
7. The method of claim 1, wherein the data block comprises 2 x 2 pixel units, and the noise reduction point is a pixel unit at a lower right corner of the current data block.
8. An image denoising method according to claim 1, wherein a is equal to 15.
9. The method of claim 1, wherein the image to be denoised is a bayer image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910290332.0A CN110097511B (en) | 2019-04-11 | 2019-04-11 | Image noise reduction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910290332.0A CN110097511B (en) | 2019-04-11 | 2019-04-11 | Image noise reduction method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110097511A CN110097511A (en) | 2019-08-06 |
| CN110097511B true CN110097511B (en) | 2021-06-15 |
Family
ID=67444763
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910290332.0A Active CN110097511B (en) | 2019-04-11 | 2019-04-11 | Image noise reduction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110097511B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102663702A (en) * | 2012-04-20 | 2012-09-12 | 西安电子科技大学 | Natural image denoising method based on regional division |
| CN103886553A (en) * | 2014-03-10 | 2014-06-25 | 广东威创视讯科技股份有限公司 | Method and system for non-local average value denoising of image |
| CN103927726A (en) * | 2014-04-23 | 2014-07-16 | 浙江宇视科技有限公司 | Image noise reduction device |
| CN104036465A (en) * | 2014-06-17 | 2014-09-10 | 南京邮电大学 | Edge detection based on self-adaptive nonlocal mean denoising method |
| CN107945119A (en) * | 2017-11-02 | 2018-04-20 | 天津大学 | Correlated noise method of estimation in image based on bayer-pattern |
| CN108109123A (en) * | 2017-12-21 | 2018-06-01 | 成都微光集电科技有限公司 | A kind of image de-noising method |
-
2019
- 2019-04-11 CN CN201910290332.0A patent/CN110097511B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102663702A (en) * | 2012-04-20 | 2012-09-12 | 西安电子科技大学 | Natural image denoising method based on regional division |
| CN103886553A (en) * | 2014-03-10 | 2014-06-25 | 广东威创视讯科技股份有限公司 | Method and system for non-local average value denoising of image |
| CN103927726A (en) * | 2014-04-23 | 2014-07-16 | 浙江宇视科技有限公司 | Image noise reduction device |
| CN104036465A (en) * | 2014-06-17 | 2014-09-10 | 南京邮电大学 | Edge detection based on self-adaptive nonlocal mean denoising method |
| CN107945119A (en) * | 2017-11-02 | 2018-04-20 | 天津大学 | Correlated noise method of estimation in image based on bayer-pattern |
| CN108109123A (en) * | 2017-12-21 | 2018-06-01 | 成都微光集电科技有限公司 | A kind of image de-noising method |
Non-Patent Citations (2)
| Title |
|---|
| Integrated Noise Modeling for Image Sensor Using Bayer Domain Images;Yeul-Min Baek 等;《Springer-Verlag Berlin Heidelberg 2009》;20091231;第413-424页 * |
| 基于非局部信息的医学图像降噪技术综述;段隆焱 等;《计算机应用研究》;20130331;第667-670页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110097511A (en) | 2019-08-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US6856704B1 (en) | Method for enhancing a digital image based upon pixel color | |
| JP4804660B2 (en) | Noise reduction method using color information for digital image processing | |
| KR101194481B1 (en) | Adjusting digital image exposure and tone scale | |
| JP4858609B2 (en) | Noise reduction device, noise reduction method, and noise reduction program | |
| JP4870617B2 (en) | Image data automatic mapping method and image processing device | |
| US9495582B2 (en) | Digital makeup | |
| CN108205804A (en) | Image processing method, device and electronic equipment | |
| CN111784605B (en) | Image noise reduction method based on region guidance, computer device and computer readable storage medium | |
| CN109978774B (en) | Denoising fusion method and device for multi-frame continuous equal exposure images | |
| JP2007310887A (en) | Automatic mapping method of image data, and image processing device | |
| KR20040044556A (en) | Image processing method, apparatus, and program | |
| JP2001216512A (en) | Method for improving digital picture having noise- dependent control of texture | |
| WO2009002111A1 (en) | Apparatus and method for automatically computing gamma correction curve | |
| CN111260580B (en) | Image denoising method, computer device and computer readable storage medium | |
| JP2008511048A (en) | Image processing method and computer software for image processing | |
| KR20070078463A (en) | Apparatus and method for reducing noise from image sensor | |
| Kanmani et al. | Swarm intelligent based contrast enhancement algorithm with improved visual perception for color images | |
| CN109903275B (en) | Fermented grain mildewing area detection method based on self-adaptive multi-scale filtering and histogram comparison | |
| Kinoshita et al. | Hue-correction scheme considering CIEDE2000 for color-image enhancement including deep-learning-based algorithms | |
| JPH0630308A (en) | Picture data processing unit | |
| CN110097511B (en) | Image noise reduction method | |
| JP5327766B2 (en) | Memory color correction in digital images | |
| CN115809966A (en) | Low-illumination image enhancement method and system | |
| CN114612344A (en) | Image sharpening device and method | |
| CN110473295B (en) | Method and equipment for carrying out beautifying treatment based on three-dimensional face model |
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 |