CN106228516B - A kind of real-time U.S. face method, apparatus of high naturalness - Google Patents
A kind of real-time U.S. face method, apparatus of high naturalness Download PDFInfo
- Publication number
- CN106228516B CN106228516B CN201610556899.4A CN201610556899A CN106228516B CN 106228516 B CN106228516 B CN 106228516B CN 201610556899 A CN201610556899 A CN 201610556899A CN 106228516 B CN106228516 B CN 106228516B
- Authority
- CN
- China
- Prior art keywords
- image
- mill skin
- skin
- mill
- sheen
- 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 abstract description 22
- 238000001914 filtration Methods 0.000 claims abstract description 34
- 230000002708 enhancing effect Effects 0.000 claims abstract description 15
- 239000000203 mixture Substances 0.000 claims description 29
- 230000003796 beauty Effects 0.000 claims description 13
- 230000002087 whitening effect Effects 0.000 abstract description 11
- 230000000694 effects Effects 0.000 abstract description 4
- 238000004364 calculation method Methods 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 6
- 230000002146 bilateral effect Effects 0.000 description 4
- 230000001965 increasing effect Effects 0.000 description 3
- 230000001815 facial effect Effects 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 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
- G06T5/00—Image enhancement or restoration
- G06T5/77—Retouching; Inpainting; Scratch removal
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20172—Image enhancement details
- G06T2207/20182—Noise reduction or smoothing in the temporal domain; Spatio-temporal filtering
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of real-time U.S. face method, apparatus of high naturalness, method includes: single channel high contrast image mill pico farad: this method also effectively reduces the noise jamming in other channels while reducing calculation amount using the least channel of noise is interfered in tri- channels r, g, b；Pachylosis information is first obtained by calculating single channel high contrast enhancing image, the directly distortion of the skin caused by image filtering is can be avoided in this way, improves the naturalness of skin；And figure layer mixes stacking method: after mill skin figure layer is calculated, present invention employs the method that three kinds of colour filter, sheen, blast mixed modes are superimposed, achieving the effect that the colour of skin grinds skin whitening, and maintain the naturalness of picture.The device of the invention makes one that face skin is allowed to keep grinding skin trace naturally, being greatly lowered while face skin whitening is smooth, face corner angle and shade is made to go out reduction degree height in mobile intelligent terminal.
Description
Technical field
The present invention relates to video, pattern U.S. face technical field, in particular to real-time beauty Yan Fangfa, the dress of a kind of high naturalness
It sets.
Background technique
Present mobile terminal, other than carrying out U.S. face when taking pictures, when video record and Video chat,
It is also required to U.S. face technology.The current user of requirement due to to(for) video U.S. face is higher and higher, and requirement of the user to U.S. face effect removes
Outside basic mill skin whitening, can also have to the naturalness of U.S. face required.So oneself skin not being intended merely in video is very
It is good, but it would also be desirable to be able to will become apparent from being ground skin (by using the tools such as figure layer, masking-out, channel, filter in PS software in addition to
Spot, the flaw or variegated outer that parts of skin can be eliminated to the personage in picture, can also grind skin for character facial, can
So that character facial is finer and smoother, profile is more clear) trace.Skin nature can be retained again while skin is soft
Texture.
In the prior art, the real-time U.S. face algorithm in mobile device is mostly based on spatial filtering algorithms, for example, Gauss
Filtering or bilateral filtering.Wherein gaussian filtering is the usual way of image noise reduction, and algorithm is by gray value around pixel
Weighted average calculate newly gray scale it, algorithm quickly but it is very poor to the reservation of image detail.Bilateral filtering is a kind of combination figure
A kind of filtering mode of image space proximity and pixel value similarity, due to consideration that the grey similarity of surrounding pixel point, this
Kind filtering mode can retain the details at face edge while even skin, but due to the details of reservation and the part filtered
Contrast visually causes result to seem that some are unnatural, such as occurs unnatural edge in skin boundary, in skin
Smooth region also has skin by honed trace.
Summary of the invention
The technical problem to be solved by the present invention is to make one to allow people while face skin whitening is smooth in mobile intelligent terminal
Face skin keeps grinding skin trace naturally, being greatly lowered, and face corner angle and shade is made to go out reduction degree height.
Above-mentioned technical problem is solved, the present invention provides a kind of real-time beauty Yan Fangfa of high naturalness, include the following steps:
Establish image filtering convolution mask；
Original image is inputted, the least channel of noise in the original image is found out and carries out gaussian filtering, obtain ash
Spend image；
Single channel high contrast image is obtained according to gray level image, and high contrast enhancing image is calculated；
Mill skin adjusting parameter is calculated by the original image, in conjunction with the mill skin adjusting parameter, high contrast enhancing figure
Picture and original image obtain mill skin image；
The mill skin image is handled, mill skin colour filter image is obtained and grinds skin sheen image；
According to the mill skin colour filter image and mill skin sheen image, the setting of mill skin image blend parameter is carried out；
Colour filter mixed image is obtained by mill skin image blend parameter and mill skin colour filter image based on the original image；
Blast mixing is obtained by mill skin image blend parameter and the colour filter mixed image based on mill skin intensified image
Image；
Sheen Mixed adjustment parameter based on setting is obtained by the mill skin sheen image and the blast mixed image
Image after to processing.
Further, described image filtering convolution mask is four layers of annulus template.
Further, pixel radius of the four layers of annulus in four layers of annulus template apart from the center of circle is respectively from small to large
2 pixels, 4 pixels, 8 pixels, 16 pixels.
Further, (I, a) * 0.1, wherein a is mill skin adjusting parameter to the mill skin adjusting parameter S=1.0+pow, and I is
The gray scale of original image.
Further, the mill skin image blend parameter M=pow (I, m), wherein m is mill skin Mixed adjustment parameter.
Further, the sheen Mixed adjustment parameter L, it is proportional with the variation tendency of sheen degree of mixing.
Further, the least channel of noise is the green channel in tri- channels r, g, b in the original image.
The present invention also provides a kind of real-time U.S. face devices of high naturalness, comprising: single channel high contrast image grinds skin list
Member and figure layer mix superpositing unit,
The single channel high contrast image grinds skin unit, to establish image filtering convolution mask；Original image is inputted, is looked for
The least channel of noise carries out gaussian filtering in the original image out, obtains gray level image；It is obtained according to gray level image
Single channel high contrast image, and high contrast enhancing image is calculated；Mill skin adjusting parameter is calculated by the original image, then
In conjunction with the mill skin adjusting parameter, high contrast enhancing image and original image, mill skin image is obtained；
The figure layer mixes superpositing unit, to handle the mill skin image, obtains mill skin colour filter image and mill
Skin sheen image；According to the mill skin colour filter image and mill skin sheen image, the setting of mill skin image blend parameter is carried out；It is based on
The original image obtains colour filter mixed image by mill skin image blend parameter and mill skin colour filter image；Based on mill skin blast
Image obtains blast mixed image by mill skin image blend parameter and the colour filter mixed image；Sheen based on setting is mixed
Close adjusting parameter, by the mill skin sheen image and the blast mixed image, image after being handled.
Further, the intelligent terminal include camera, mobile phone, plate, notebook, desktop computer can photograph set
It is standby
Beneficial effects of the present invention:
1) the real-time beauty Yan Fangfa of the high naturalness in the present invention, scheme compared with the prior art, with making one face
Skin whitening allows face skin to keep naturally, mill skin trace is greatly lowered while smooth, face corner angle and shade is made to go out to restore
Spend high beautiful Yan Xiaoguo.This method, which can reduce, directly carries out the unnatural degree of bilateral filtering bring image to image, is scheming
Layer hybrid mode on use colour filter, sheen, the method that three kinds of modes of blast are superimposed, to face skin mill skin and whitening it is same
When ensure that the naturalness of image, face edge in picture will not be made to be mutated, natural details line on skin will not be lost
Reason, so that skin beauty and true to nature.
2) due to establishing image filtering convolution mask；Original image is inputted, it is least to find out noise in the original image
One channel carries out gaussian filtering, obtains gray level image；Single channel high contrast image is obtained according to gray level image, and is calculated
High contrast enhances image；Mill skin adjusting parameter is calculated by the original image, in conjunction with the mill skin adjusting parameter, high contrast
Enhance image and original image, obtains mill skin image；It is not in the present invention space filtering directly to be carried out to image, but adopt
With by calculating single channel high contrast image, to weaken pachylosis part in image by adjusting mill skin coefficient.
3) due to handling the mill skin image, obtaining mill skin colour filter image and grinding skin sheen image；According to described
Skin colour filter image and mill skin sheen image are ground, the setting of mill skin image blend parameter is carried out；Based on the original image, pass through mill
Skin image blend parameter and mill skin colour filter image, obtain colour filter mixed image；It is mixed by mill skin image based on mill skin intensified image
Parameter and the colour filter mixed image are closed, blast mixed image is obtained；Sheen Mixed adjustment parameter based on setting, by described
Grind skin sheen image and the blast mixed image, image after being handled.Colour filter is used in figure layer hybrid mode, it is soft
Light, the method that three kinds of modes of blast are superimposed ensure that the naturalness of image while grinding skin and whitening to face skin.
4) in the present invention high naturalness real-time U.S. face device, single channel high contrast image mill is provided on intelligent terminal
Skin unit, the single channel high contrast image mill pico farad used in the unit: this method is used to interfere in tri- channels r, g, b and be made an uproar
The least channel of point, also effectively reduces the noise jamming in other channels while reducing calculation amount；By calculating single channel
High contrast enhances image first to obtain pachylosis information, can be avoided the directly distortion of the skin caused by image filtering in this way,
Improve the naturalness of skin.
5) in the present invention high naturalness real-time U.S. face device, figure layer mixing superpositing unit is provided on intelligent terminal
Figure, the layer mixing stacking method used in the unit: after mill skin figure layer is calculated, present invention employs colour filter, sheen, increasings
The method of bright three kinds of mixed modes superposition, achievees the effect that the colour of skin grinds skin whitening, and maintain the naturalness of picture.
Detailed description of the invention
Fig. 1 is the real-time U.S. face method flow schematic diagram of the high naturalness in one embodiment of the invention.
Fig. 2 is the real-time U.S. face apparatus structure schematic diagram of the high naturalness in one embodiment of the invention.
Fig. 3 is the real-time U.S. face apparatus structure schematic diagram of the high naturalness in another embodiment of the present invention.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with specific embodiment, and reference
Attached drawing, the present invention is described in more detail.
It is the real-time U.S. face method flow schematic diagram of the high naturalness in one embodiment of the invention as shown in Figure 1.
The real-time beauty Yan Fangfa of the high naturalness of one of one embodiment of the invention, includes the following steps:
Step S100 establishes image filtering convolution mask；In some embodiments, described image filtering convolution mask is four
Layer annulus template, pixel radius of the four layers of annulus apart from the center of circle in four layers of annulus template be respectively from small to large 2 pixels,
4 pixels, 8 pixels, 16 pixels.
Step S101 inputs original image, finds out the least channel of noise in the original image and carries out Gauss filter
Wave obtains gray level image；Original image is denoted as Image0, the least channel of original image noise is found out, to the channel
Gaussian filtering is carried out, a single pass gray level image is obtained, is denoted as Image1
Step S102 obtains single channel high contrast image according to gray level image, and high contrast enhancing image is calculated；Meter
Single channel high contrast image Image2=Image0-Image1+0.5 is calculated, further, calculating high contrast enhances image Image3,
If Image3 < 0.5, Image3=2.0*Image2*Image2；
If Image3 > 0.5, Image3=1.0- (2.0* (1.0-Image2) * (1.0-Image2))；
Step S103 calculates mill skin adjusting parameter by the original image, anti-in conjunction with the mill skin adjusting parameter, height
Difference enhancing image and original image obtain mill skin image；In some embodiments, the gray scale I of original image Image0 is calculated, if
Surely grinding skin coefficient S=1.0+pow, (I, a) * 0.1, wherein a is mill skin adjusting parameter, and the smaller mill skin smoothness of value is higher；It calculates
Obtain the mill skin image Image4:Image4=Image0*S-Image3* (S-1.0).
Step S104 handles the mill skin image, obtains mill skin colour filter image and grinds skin sheen image；To described
Mill skin image is handled including but not limited to, calculating mill skin intensified image Image6:Image6=max (Image0,
Image4)；
Calculate mill skin colour filter image Image5:Image5=1.0- (1.0-Image0) * (1.0-Image4)；
Calculate the mill skin sheen image Image7:Image7=2.0*Image0*Image4+Image0*Image0-
2.0*Image0*Image0*Image4；
Step S105 carries out setting for mill skin image blend parameter according to the mill skin colour filter image and mill skin sheen image
It is fixed；It grinds skin image blend parameter M=pow (I, m), wherein m is mill skin Mixed adjustment parameter, and the smaller mill skin degree of mixing of value is got over
It is high；
Step S106 is based on the original image, by mill skin image blend parameter and mill skin colour filter image, obtains colour filter
Mixed image；Calculate colour filter mixed image Image8:Image8=M*Image5+ (1.0-M) * Image0；
Step S107 is based on mill skin intensified image and is obtained by mill skin image blend parameter and the colour filter mixed image
Blast mixed image；Calculate blast mixed image Image9:Image9=M*Image6+ (1.0-M) * Image8；
Sheen Mixed adjustment parameter of the step S108 based on setting is mixed by the mill skin sheen image and the blast
Image, image after being handled.Obtain final result image ImageR: the setting bigger sheen of sheen Mixed adjustment parameter L, L is mixed
It is right higher, ImageR=L*Image7+ (1.0-L) * Image9.
As preferred in the present embodiment, (I, a) * 0.1, wherein a is mill skin to the mill skin adjusting parameter S=1.0+pow
Adjusting parameter, I are the gray scale of original image.
As preferred in the present embodiment, the mill skin image blend parameter M=pow (I, m), wherein m is mill skin mixing
Adjusting parameter.
As preferred in the present embodiment, the sheen Mixed adjustment parameter L, the variation tendency with sheen degree of mixing is in just
Than.
As preferred in the present embodiment, the least channel of noise is tri- channels r, g, b in the original image
In green channel.
The real-time beauty Yan Fangfa of high naturalness in the present embodiment, scheme compared with the prior art, having makes one face
Skin whitening allows face skin to keep naturally, mill skin trace is greatly lowered while smooth, face corner angle and shade is made to go out to restore
Spend high beautiful Yan Xiaoguo.This method, which can reduce, directly carries out the unnatural degree of bilateral filtering bring image to image, is scheming
Layer hybrid mode on use colour filter, sheen, the method that three kinds of modes of blast are superimposed, to face skin mill skin and whitening it is same
When ensure that the naturalness of image, face edge in picture will not be made to be mutated, natural details line on skin will not be lost
Reason, so that skin beauty and true to nature.
The principle of the present invention:
1. make image filtering convolution template: the volume machine template that this method uses for four layers of annulus template, four layers of annulus away from
Pixel radius from the center of circle is respectively 2 pixels, 4 pixels, 8 pixels, 16 pixels from small to large；
2. single channel gaussian filtering: original image being denoted as Image0, finds out the least channel of original image noise
(for example being usually that green channel noise is minimum on iphone camera), carries out gaussian filtering to the channel, obtains a width single channel
Gray level image, be denoted as Image1；
3. calculating single channel high contrast image Image2:Image2=Image0-Image1+0.5；
4. calculating high contrast enhances image Image3:
If Image3 < 0.5, Image3=2.0*Image2*Image2；
If Image3 > 0.5, Image3=1.0- (2.0* (1.0-Image2) * (1.0-Image2))；
5. calculating mill skin adjusting parameter S: calculating the gray scale I of original image Image0, setting mill skin coefficient S=1.0+pow
(I, a) * 0.1, wherein a is mill skin adjusting parameter, and the smaller mill skin smoothness of value is higher；
6. calculating mill skin image Image4:Image4=Image0*S-Image3* (S-1.0)；
7. calculating mill skin colour filter image Image5:Image5=1.0- (1.0-Image0) * (1.0-Image4)；
8. calculating mill skin intensified image Image6:Image6=max (Image0, Image4)；
9. calculating mill skin sheen image Image7:Image7=2.0*Image0*Image4+Image0*Image0-2.0*
Image0*Image0*Image4；
10. setting mill skin image blend parameter M:M=pow (I, m), wherein m is mill skin Mixed adjustment parameter, and value is smaller
It is higher to grind skin degree of mixing；
11. calculating colour filter mixed image Image8:Image8=M*Image5+ (1.0-M) * Image0；
12. calculating blast mixed image Image9:Image9=M*Image6+ (1.0-M) * Image8；
13. obtaining final result image ImageR: the setting bigger sheen degree of mixing of sheen Mixed adjustment parameter L, L is higher,
ImageR=L*Image7+ (1.0-L) * Image9.
It is the real-time U.S. face apparatus structure schematic diagram of the high naturalness in one embodiment of the invention as shown in Figure 2.
The real-time U.S. face device 20 of the high naturalness of one of the present embodiment, is set to intelligent terminal, comprising: single channel is high
Contrast image grinds skin unit 200 and figure layer mixes superpositing unit 201,
The single channel high contrast image grinds skin unit 200, to establish image filtering convolution mask；Input original graph
Picture finds out the least channel of noise in the original image and carries out gaussian filtering, obtains gray level image；According to gray level image
Single channel high contrast image is obtained, and high contrast enhancing image is calculated；
Mill skin adjusting parameter is calculated by the original image, in conjunction with the mill skin adjusting parameter, high contrast enhancing figure
Picture and original image obtain mill skin image；
The figure layer mixes superpositing unit 201, handling the mill skin image, obtain mill skin colour filter image with
Grind skin sheen image；According to the mill skin colour filter image and mill skin sheen image, the setting of mill skin image blend parameter is carried out；Base
Colour filter mixed image is obtained by mill skin image blend parameter and mill skin colour filter image in the original image；Increased based on mill skin
Bright image obtains blast mixed image by mill skin image blend parameter and the colour filter mixed image；Sheen based on setting
Mixed adjustment parameter, by the mill skin sheen image and the blast mixed image, image after being handled.
The real-time U.S. face device of high naturalness, is provided with single channel high contrast figure on intelligent terminal in the present embodiment
As mill skin unit 200, the single channel high contrast image used in the unit grinds pico farad: this method uses tri- channels r, g, b
The least channel of middle interference noise, also effectively reduces the noise jamming in other channels while reducing calculation amount；Pass through meter
Single channel high contrast enhancing image is calculated first to obtain pachylosis information, can be avoided the directly skin caused by image filtering in this way
Skin distortion, improves the naturalness of skin.
The real-time U.S. face device of high naturalness in the present embodiment, is provided with figure layer mixing superpositing unit on intelligent terminal
Figure 20 1, the layer mixing stacking method used in the unit: after mill skin figure layer is calculated, present invention employs colour filters, soft
The method that three kinds of light, blast mixed modes are superimposed achievees the effect that the colour of skin grinds skin whitening, and maintains the naturalness of picture.
It is the real-time U.S. face apparatus structure schematic diagram of the high naturalness in another embodiment of the present invention as shown in Figure 3.
The real-time U.S. face device 20 of the high naturalness of one of the present embodiment, is set to intelligent terminal, comprising: single channel is high
Contrast image grinds skin unit 200 and figure layer mixes superpositing unit 201,
The single channel high contrast image grinds skin unit 200, to establish image filtering convolution mask；Input original graph
Picture finds out the least channel of noise in the original image and carries out gaussian filtering, obtains gray level image；According to gray level image
Single channel high contrast image is obtained, and high contrast enhancing image is calculated；Mill skin adjustment ginseng is calculated by the original image
Number obtains mill skin image in conjunction with the mill skin adjusting parameter, high contrast enhancing image and original image；
The figure layer mixes superpositing unit 201, handling the mill skin image, obtain mill skin colour filter image with
Grind skin sheen image；According to the mill skin colour filter image and mill skin sheen image, the setting of mill skin image blend parameter is carried out；Base
Colour filter mixed image is obtained by mill skin image blend parameter and mill skin colour filter image in the original image；Increased based on mill skin
Bright image obtains blast mixed image by mill skin image blend parameter and the colour filter mixed image；Sheen based on setting
Mixed adjustment parameter, by the mill skin sheen image and the blast mixed image, image after being handled.Preferably, institute
State the equipment that can be photographed that intelligent terminal includes camera (not shown), mobile phone 40, plate 30, notebook 60, desktop computer 30.
More than it should be understood by those ordinary skilled in the art that:, described is only specific embodiments of the present invention, and
It is not used in the limitation present invention, all within the spirits and principles of the present invention, any modification, equivalent substitution, improvement and etc. done,
It should be included within protection scope of the present invention.
Claims (8)
1. a kind of real-time beauty Yan Fangfa of high naturalness, which comprises the steps of:
Establish image filtering convolution mask；
Original image is inputted, the least channel of noise in the original image is found out and carries out gaussian filtering, obtain grayscale image；
Single channel high contrast image is obtained according to gray level image, and high contrast enhancing image is calculated；
Mill skin coefficient is calculated by the original image, enhances image and original image in conjunction with the mill skin coefficient, high contrast,
Mill skin image is obtained, (I, a) * 0.1, wherein a is mill skin adjusting parameter to the mill skin coefficient S=1.0+pow, and I is original image
Gray scale；
The mill skin image is handled, mill skin colour filter image is obtained and grinds skin sheen image；
According to the mill skin colour filter image and mill skin sheen image, the setting of mill skin image blend parameter is carried out；
Colour filter mixed image is obtained by mill skin image blend parameter and mill skin colour filter image based on the original image；
Blast mixed image is obtained by mill skin image blend parameter and the colour filter mixed image based on mill skin intensified image；
Sheen Mixed adjustment parameter based on setting is obtained everywhere by the mill skin sheen image and the blast mixed image
Image after reason.
2. real-time beauty Yan Fangfa according to claim 1, which is characterized in that it is four layers of circle that described image, which filters convolution mask,
Ring moulds version.
3. real-time beauty Yan Fangfa according to claim 2, which is characterized in that four layers of annulus in four layers of annulus template
Pixel radius apart from the center of circle is respectively 2 pixels, 4 pixels, 8 pixels, 16 pixels from small to large.
4. real-time beauty Yan Fangfa according to claim 1, which is characterized in that the mill skin image blend parameter M=pow
(I, m), wherein m is mill skin Mixed adjustment parameter.
5. real-time beauty Yan Fangfa according to claim 1, which is characterized in that the sheen Mixed adjustment parameter L, with sheen
The variation tendency of degree of mixing is proportional.
6. real-time beauty Yan Fangfa according to claim 1, which is characterized in that least one of noise in the original image
Channel is the green channel in tri- channels r, g, b.
7. a kind of real-time U.S. face device of high naturalness, is set to intelligent terminal characterized by comprising single channel high contrast
Image grinds skin unit and figure layer mixes superpositing unit,
The single channel high contrast image grinds skin unit, to establish image filtering convolution mask；Original image is inputted, institute is found out
It states the least channel of noise in original image and carries out gaussian filtering, obtain gray level image；Single-pass is obtained according to gray level image
Road high contrast image, and high contrast enhancing image is calculated；Mill skin coefficient is calculated by the original image, in conjunction with described
Skin coefficient, high contrast enhancing image and original image are ground, mill skin image, the mill skin coefficient S=1.0+pow (I, a) * are obtained
0.1, wherein a is mill skin adjusting parameter, and I is the gray scale of original image；
The figure layer mixes superpositing unit, to handle the mill skin image, obtains mill skin colour filter image and mill skin is soft
Light image；According to the mill skin colour filter image and mill skin sheen image, the setting of mill skin image blend parameter is carried out；Based on described
Original image obtains colour filter mixed image by mill skin image blend parameter and mill skin colour filter image；Based on mill skin blast figure
Picture obtains blast mixed image by mill skin image blend parameter and the colour filter mixed image；Sheen mixing based on setting
Adjusting parameter, by the mill skin sheen image and the blast mixed image, image after being handled.
8. real-time U.S. face device according to claim 7, which is characterized in that the intelligent terminal includes camera, mobile phone, puts down
The equipment that can be photographed of plate, notebook, desktop computer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610556899.4A CN106228516B (en) | 2016-07-14 | 2016-07-14 | A kind of real-time U.S. face method, apparatus of high naturalness |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610556899.4A CN106228516B (en) | 2016-07-14 | 2016-07-14 | A kind of real-time U.S. face method, apparatus of high naturalness |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106228516A CN106228516A (en) | 2016-12-14 |
| CN106228516B true CN106228516B (en) | 2019-07-19 |
Family
ID=57520390
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610556899.4A Active CN106228516B (en) | 2016-07-14 | 2016-07-14 | A kind of real-time U.S. face method, apparatus of high naturalness |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106228516B (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106600556A (en) * | 2016-12-16 | 2017-04-26 | 合网络技术（北京）有限公司 | Image processing method and apparatus |
| CN107730445B (en) * | 2017-10-31 | 2022-02-18 | Oppo广东移动通信有限公司 | Image processing method, image processing apparatus, storage medium, and electronic device |
| CN109146823A (en) * | 2018-09-25 | 2019-01-04 | 广州华多网络科技有限公司 | Image processing method, device, equipment and the storage medium retained based on high contrast |
| CN109377454A (en) * | 2018-09-25 | 2019-02-22 | 广州华多网络科技有限公司 | A kind of image processing method, device, equipment, storage medium and live broadcasting method |
| CN110070494B (en) * | 2018-12-21 | 2021-09-17 | 北京字节跳动网络技术有限公司 | Image processing method and device and electronic equipment |
| CN109741272A (en) * | 2018-12-25 | 2019-05-10 | 维沃移动通信有限公司 | A kind of image processing method and mobile terminal |
| CN110417990B (en) * | 2019-03-25 | 2020-07-24 | 浙江麦知网络科技有限公司 | APP starting system based on target analysis |
| CN111028180B (en) * | 2019-12-23 | 2021-11-09 | 腾讯科技（深圳）有限公司 | Image processing method, video processing method and corresponding devices |
| CN112215768A (en) * | 2020-09-30 | 2021-01-12 | 广州虎牙科技有限公司 | Image definition improving method and device, electronic equipment and readable storage medium |
| CN112233036B (en) * | 2020-10-21 | 2021-06-08 | 哈尔滨市科佳通用机电股份有限公司 | Image enhancement method and system based on high contrast retention algorithm |
| CN113628196A (en) * | 2021-08-16 | 2021-11-09 | 广东艾檬电子科技有限公司 | Image content extraction method, device, terminal and storage medium |
| CN114240743B (en) * | 2021-12-15 | 2024-06-25 | 重庆米弘科技有限公司 | Skin beautifying method based on high-contrast skin grinding face image |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6108455A (en) * | 1998-05-29 | 2000-08-22 | Stmicroelectronics, Inc. | Non-linear image filter for filtering noise |
| CN104504661A (en) * | 2014-11-27 | 2015-04-08 | 厦门美图之家科技有限公司 | Method for processing image to be in black and white effect |
| CN105243371A (en) * | 2015-10-23 | 2016-01-13 | 厦门美图之家科技有限公司 | Human face beauty degree detection method and system and shooting terminal |
| CN105608677A (en) * | 2015-12-28 | 2016-05-25 | 成都品果科技有限公司 | Image skin color beautifying method and system under any lighting conditions |
| CN105741231A (en) * | 2016-02-02 | 2016-07-06 | 深圳中博网络技术有限公司 | Skin beautifying processing method and device of image |
-
2016
- 2016-07-14 CN CN201610556899.4A patent/CN106228516B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6108455A (en) * | 1998-05-29 | 2000-08-22 | Stmicroelectronics, Inc. | Non-linear image filter for filtering noise |
| CN104504661A (en) * | 2014-11-27 | 2015-04-08 | 厦门美图之家科技有限公司 | Method for processing image to be in black and white effect |
| CN105243371A (en) * | 2015-10-23 | 2016-01-13 | 厦门美图之家科技有限公司 | Human face beauty degree detection method and system and shooting terminal |
| CN105608677A (en) * | 2015-12-28 | 2016-05-25 | 成都品果科技有限公司 | Image skin color beautifying method and system under any lighting conditions |
| CN105741231A (en) * | 2016-02-02 | 2016-07-06 | 深圳中博网络技术有限公司 | Skin beautifying processing method and device of image |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106228516A (en) | 2016-12-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106228516B (en) | A kind of real-time U.S. face method, apparatus of high naturalness | |
| Li et al. | Edge-preserving decomposition-based single image haze removal | |
| CN106780311B (en) | Rapid face image beautifying method combining skin roughness | |
| US9495582B2 (en) | Digital makeup | |
| US9626751B2 (en) | Method and system for analog/digital image simplification and stylization | |
| CN106971165B (en) | A kind of implementation method and device of filter | |
| Li et al. | Visual-salience-based tone mapping for high dynamic range images | |
| CN107369133B (en) | Face image beautifying method and device | |
| US7768529B2 (en) | Method and system for image editing | |
| CN108205804A (en) | Image processing method, device and electronic equipment | |
| CN104899905B (en) | Face image processing process and device | |
| CN106326823B (en) | Method and system for obtaining head portrait in picture | |
| Hanumantharaju et al. | Color image enhancement using multiscale retinex with modified color restoration technique | |
| CN105023244B (en) | Method and device for beautifying human skin color in image and adjusting human skin color brightness | |
| CN107871303A (en) | A kind of image processing method and device | |
| CN104282002A (en) | Quick digital image beautifying method | |
| CN107516302A (en) | A kind of method of the mixed image enhancing based on OpenCV | |
| CN105023252A (en) | Method and system for enhancement processing of beautified image and shooting terminal | |
| CN111861869B (en) | Image processing method and device for beautifying human images and preventing background distortion | |
| CN103035019B (en) | Image processing method and device | |
| CN112215776A (en) | Portrait buffing method, electronic device and computer readable storage medium | |
| CN116168091A (en) | Image processing method, apparatus, computer device and computer program product | |
| CN113177878B (en) | Method and device for realizing American cartoon style filter effect based on image transformation | |
| CN108230271A (en) | Cosmetic method on face foundation cream in a kind of digital picture based on Face datection and facial feature localization | |
| CN113673270B (en) | Image processing method and device, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20180621 Address after: 100041 B-0035, 2 floor, 3 building, 30 Shixing street, Shijingshan District, Beijing. Applicant after: BEIJING BYTEDANCE NETWORK TECHNOLOGY Co.,Ltd. Address before: 518000 Room 201, building A, No. 1, Qian Wan Road, Qianhai Shenzhen Hong Kong cooperation zone, Shenzhen, Guangdong (Shenzhen Qianhai business secretary Co., Ltd.) Applicant before: LIANMENG TECHNOLOGY (SHENZHEN) CO.,LTD. |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP01 | Change in the name or title of a patent holder |
Address after: 100041 B-0035, 2 floor, 3 building, 30 Shixing street, Shijingshan District, Beijing. Patentee after: Tiktok vision (Beijing) Co.,Ltd. Address before: 100041 B-0035, 2 floor, 3 building, 30 Shixing street, Shijingshan District, Beijing. Patentee before: BEIJING BYTEDANCE NETWORK TECHNOLOGY Co.,Ltd. Address after: 100041 B-0035, 2 floor, 3 building, 30 Shixing street, Shijingshan District, Beijing. Patentee after: Douyin Vision Co.,Ltd. Address before: 100041 B-0035, 2 floor, 3 building, 30 Shixing street, Shijingshan District, Beijing. Patentee before: Tiktok vision (Beijing) Co.,Ltd. |
|
| CP01 | Change in the name or title of a patent holder |