CN107623845B - A kind of image processing method and device based on priori knowledge - Google Patents
A kind of image processing method and device based on priori knowledge Download PDFInfo
- Publication number
- CN107623845B CN107623845B CN201610549868.6A CN201610549868A CN107623845B CN 107623845 B CN107623845 B CN 107623845B CN 201610549868 A CN201610549868 A CN 201610549868A CN 107623845 B CN107623845 B CN 107623845B
- Authority
- CN
- China
- Prior art keywords
- skin color
- area
- image
- pixel
- adjustment
- 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
Landscapes
- Image Processing (AREA)
- Facsimile Image Signal Circuits (AREA)
Abstract
The present invention provides a kind of image processing method and device based on priori knowledge, and the image processing method based on priori knowledge includes: the first area of skin color for obtaining the image of chrominance space；According to the chrominance information and preset skin color range of each pixel in the first area of skin color, the confidence value of the first area of skin color is determined；When the confidence value of the first area of skin color is not more than preset threshold, according to the position of each pixel in the first area of skin color and preset skin color range, adjust the position of each pixel in the first area of skin color, obtain the second area of skin color, wherein, the relative position between each pixel in the second area of skin color is corresponding with the relative position between each pixel in the first area of skin color；According to the first area of skin color and the second area of skin color, the adjusted value of the image of chrominance space is determined；The adjustment of image colour cast is carried out according to the adjusted value of the image of chrominance space, the image after obtaining white balance.It realizes the reduction image color difference being simple and efficient, improves the purpose of image rectification effect.
Description
Technical field
The present invention relates to technical field of image processing, more particularly to a kind of image processing method based on priori knowledge and
Device.
Background technique
Automatic white balance (Auto White Balance) is a very important ring in digital picture pretreatment system
Section.So-called white balance can be understood as the photo color of white object being reduced to white.Object have respective color be because
Its surface selectively absorbs and reflects the visible light of part wavelength section, and wherein the light of reflective portion is object institute table
Existing color.When the light of light source is not natural light, and has certain color, the light that object is reflected is not only it
The color of itself, while also having the color of light source.Human eye has very strong adaptability, can in the environment of colored light sources
The color of a part of light source of automatic shield, i.e., from people perceptually for object itself color is still substantially presented.However
The sensor of digital camera does not have this color adaptation ability, it can acquire the color of light source, and photo is caused also to take equally
Color offset.In order to make photo effect close with the perception of human eye, needing removal when necessary or weaken the influence of light source color, this
It is exactly so-called white balance.And automatic white balance is that the colour cast situation of photo is automatically detected when taking pictures, and is taken measures in due course
The strong and weak ratio of each Color Channel of RGB is adjusted, to achieve the effect that white balance.
Traditional automatic white balance algorithm includes gray world method, total reflection theoretical algorithm etc., they scheme by acquisition
Characteristics of image is analyzed as certain statistical informations, thus the method for acquiring correction.However gray world algorithm is due to only to pixel
The direct average weighted of RGB component cannot carry out more intelligent processing to image, although having certain correction to white recovery,
But it loses in terms of brightness of image, and single for color, color image rectification effect not abundant is bad.Total reflection
Theoretical algorithm is bad for the image processing effect of serious colour cast, since the theory is proposed based on brightness maximum region in image
, so not being the image of white area for brightness maximum region in image, correcting algorithm can generate error.
Summary of the invention
In view of the foregoing deficiencies of prior art, the purpose of the present invention is to provide a kind of images based on priori knowledge
Processing method and processing device, the reduction image color difference for being simple and efficient improve image rectification effect.
In order to achieve the above objects and other related objects, the present invention provides a kind of image processing method based on priori knowledge
Method, comprising the following steps: obtain the first area of skin color of the image of chrominance space；According to each picture in first area of skin color
The chrominance information and preset skin color range of element, determine the confidence value of first area of skin color；In first colour of skin area
When the confidence value in domain is not more than preset threshold, according to the position of each pixel in the first area of skin color and preset skin color range,
The position for adjusting each pixel in first area of skin color, obtains the second area of skin color, wherein in second area of skin color
For part or all of pixel in the default skin color range, the second area of skin color confidence value is greater than preset threshold；It is described
The relative position between each pixel in second area of skin color is corresponding with the relative position between each pixel in first area of skin color；
According to first area of skin color and second area of skin color, the adjusted value of the image of the chrominance space is determined；According to institute
The adjusted value for stating the image of chrominance space carries out the adjustment of image colour cast, the image after obtaining white balance.
In one embodiment of the invention, the first area of skin color of the image for obtaining chrominance space includes: that will input
Image the image of chrominance space is converted to by RGB；According to the parameter information in the image of the chrominance space, determine described
The first area of skin color in the image of chrominance space.
In one embodiment of the invention, the chrominance information according to each pixel in first area of skin color and pre-
If skin color range, determine that the confidence value of first area of skin color includes: according to each picture in first area of skin color
The chrominance information and preset skin color range of element, utilize formula
Calculate the confidence value of first area of skin color；Wherein, n is the sum of all pixels in area of skin color, and i is to be not more than greater than 0
The integer of n, TiIndicate the confidence value of ith pixel, T is the confidence value of first area of skin color.
In one embodiment of the invention, the confidence value in first area of skin color is not more than preset threshold
When, according to the position of each pixel in the first area of skin color and preset skin color range, adjust each picture in first area of skin color
The position of element, obtaining the second area of skin color includes: the root when the confidence value of first area of skin color is no more than preset threshold
According to first area of skin color and the preset skin color range, the adjustment vector of first area of skin color is determined；Described
The adjustment vector of one area of skin color is for adjusting each pixel in the first area of skin color into the preset skin color range
Vector；Execute set-up procedure；Wherein, the set-up procedure includes: according to the adjustment vector of first area of skin color, adjustment
The position of each pixel in first area of skin color, and calculate the confidence value of the first area of skin color adjusted；Described in determination
Whether the confidence value of the first area of skin color adjusted is greater than preset threshold；In first area of skin color adjusted can
When certainty value is greater than preset threshold, first area of skin color adjusted is determined as the second area of skin color；In the adjustment
When the confidence value of the first area of skin color afterwards is not more than preset threshold, the is updated using first area of skin color adjusted
One area of skin color re-executes the set-up procedure.
In one embodiment of the invention, the confidence value in first area of skin color is not more than preset threshold
When, according to first area of skin color and the preset skin color range, determine the adjustment vector packet of first area of skin color
It includes: when the confidence value of first area of skin color is no more than preset threshold, according to the pixel in first area of skin color
The position of point and the preset skin color range, determine adjustment direction；According to the adjustment direction and it is default be sized, really
The adjustment vector of fixed first area of skin color.
In one embodiment of the invention, the confidence value in first area of skin color is not more than preset threshold
When, according to the position of the pixel in first area of skin color and the preset skin color range, determine adjustment direction packet
It includes: when the confidence value of first area of skin color is no more than preset threshold, the is determined in first area of skin color
One rectangle；First rectangle is that the smallest each pixel by first area of skin color of area of the first area of skin color is wrapped
It is trapped among interior rectangle；The second rectangle is determined in the preset skin color range；Second rectangle be and first square
Shape has identical breadth length ratio, the maximum rectangle of each side parallel area corresponding with first rectangle；By first rectangle
The direction of the central point central point that is directed toward second rectangle be determined as the adjustment direction.
In one embodiment of the invention, the adjusted value of the image according to the chrominance space carries out image colour cast tune
Whole, the image after obtaining white balance includes: to determine the G in rgb color space according to the adjusted value of the image of the chrominance space
The adjustment gray value in channel；It is determined in rgb color space according to the adjustment gray value in the channel G using histogram matching
The channel R adjustment gray value and channel B；According to the adjustment gray value in the channel G in the rgb color space, the tune in the channel R
Whole gray value and the adjustment gray value of channel B carry out the adjustment of image colour cast, the image of the rgb color space after obtaining white balance.
In one embodiment of the invention, parameter information in the image according to the chrominance space determines institute
Stating the first area of skin color in the image of chrominance space includes: the parameter information in image according to the chrominance space, is utilized
Parametrization complexion model method determines the first area of skin color in the image of the chrominance space.
In one embodiment of the invention, the chrominance space is YCbCr chrominance space；The preset skin color range packet
It includes: 133 < Cr <, 177,77 < Cb < 127 and 125 < Cb+0.6Cr < 190.
Further, the present invention provides a kind of image processing apparatus, comprising: processing unit, for obtaining chrominance space
First area of skin color of image；Determination unit, it is each in first area of skin color for being obtained according to the processing unit
The chrominance information of pixel and preset skin color range determine the confidence value of first area of skin color；The processing unit, also
When the confidence value of first area of skin color for determining in the determination unit is not more than preset threshold, according to the first skin
The position of each pixel and preset skin color range in color region adjust the position of each pixel in first area of skin color, obtain
Second area of skin color, wherein pixel is described in the default skin color range some or all of in second area of skin color
Second area of skin color confidence value is greater than preset threshold；Relative position between each pixel in second area of skin color with it is described
Relative position in first area of skin color between each pixel is corresponding；The determination unit is also used to according to first area of skin color
And second area of skin color that the processing unit obtains, determine the adjusted value of the image of the chrominance space；Adjustment unit,
The adjusted value of the image of the chrominance space for being determined according to the determination unit carries out the adjustment of image colour cast, obtains white flat
Image after weighing apparatus.
In one embodiment of the invention, the processing unit is turned specifically for the image that will be inputted by rgb color space
It is changed to the image of chrominance space；According to the parameter information in the image of the chrominance space, the figure of the chrominance space is determined
The first area of skin color as in.
In one embodiment of the invention, the determination unit, specifically for according to each picture in first area of skin color
The chrominance information and preset skin color range of element, utilize formula
Calculate the confidence value of first area of skin color；Wherein, n is the sum of all pixels in area of skin color, and i is to be not more than greater than 0
The integer of n, TiIndicate the confidence value of ith pixel, T is the confidence value of first area of skin color.
In one embodiment of the invention, the determination unit is also used to the confidence value in first area of skin color
When no more than preset threshold, according to first area of skin color and the preset skin color range, first colour of skin area is determined
The adjustment vector in domain；The adjustment vector of first area of skin color be for by each pixel in the first area of skin color to described pre-
If skin color range in adjust vector；The processing unit is specifically used for executing set-up procedure；Wherein, the set-up procedure
Include: the adjustment vector of first area of skin color determined according to the determination unit, adjusts in first area of skin color
The position of each pixel, and calculate the confidence value of the first area of skin color adjusted；Determine the first colour of skin adjusted area
Whether the confidence value in domain is greater than preset threshold；It is greater than preset threshold in the confidence value of first area of skin color adjusted
When, first area of skin color adjusted is determined as the second area of skin color；In first area of skin color adjusted
When confidence value is not more than preset threshold, the first area of skin color is updated using first area of skin color adjusted, is held again
The row set-up procedure.
In one embodiment of the invention, the determination unit, specifically for the confidence level in first area of skin color
When value is no more than preset threshold, according to the position of the pixel in first area of skin color and the preset skin color range,
Determine adjustment direction；According to the adjustment direction and it is default be sized, determine the adjustment vector of first area of skin color.
In one embodiment of the invention, the determination unit is not more than in the confidence value of first area of skin color
When preset threshold, the first rectangle is determined in first area of skin color；First rectangle is the face of the first area of skin color
The smallest each pixel by first area of skin color of product is enclosed in interior rectangle；In the preset skin color range really
Make the second rectangle；Second rectangle is to have identical breadth length ratio, each side and first rectangle with first rectangle
The corresponding parallel maximum rectangle of area；The central point of first rectangle is directed toward to the direction of the central point of second rectangle
It is determined as the adjustment direction.
In one embodiment of the invention, the adjustment unit, specifically for the tune according to the image of the chrominance space
Whole value determines the adjustment gray value in the channel G in rgb color space；According to the adjustment gray value in the channel G, histogram is utilized
Figure matching method determines the adjustment gray value and channel B in the channel R in rgb color space；According to the G in the rgb color space
The adjustment gray value in channel, the adjustment gray value in the channel R and the adjustment gray value of channel B carry out the adjustment of image colour cast, obtain white
The image of rgb color space after balance.
In one embodiment of the invention, the processing unit, specifically in the image according to the chrominance space
Parameter information determines the first area of skin color in the image of the chrominance space using parametrization complexion model method.
In one embodiment of the invention, the chrominance space is YCbCr chrominance space；The preset skin color range packet
It includes: 133 < Cr <, 177,77 < Cb < 127 and 125 < Cb+0.6Cr < 190.
As described above, a kind of image processing method and device based on priori knowledge of the invention, has below beneficial to effect
Fruit: image processing apparatus of the present invention gets the tune of the image of chrominance space using the first area of skin color of the image of chrominance space
Whole value, the image so as to carry out the adjustment of image colour cast according to the adjusted value of the image of chrominance space, after obtaining white balance.I.e.
To be to get the adjusted value of image according to the first area of skin color in image, and then carry out the adjustment of image colour cast in the present invention,
So that the present invention is adjusted image colour cast using relatively simple method, and accuracy is high, is simple and efficient to realize
Reduce image color difference, improves the purpose of image rectification effect.
Detailed description of the invention
Fig. 1 is shown as a kind of method based on the image processing method of priori knowledge in an embodiment provided by the invention
Schematic diagram.
Fig. 2 is shown as the first exemplary schematic diagram of the image processing method provided by the invention based on priori knowledge.
Fig. 3 is shown as exemplary second of schematic diagram of the image processing method provided by the invention based on priori knowledge.
Fig. 4 is shown as the third exemplary schematic diagram of the image processing method provided by the invention based on priori knowledge.
Fig. 5 is shown as the exemplary 4th kind of schematic diagram of the image processing method provided by the invention based on priori knowledge.
Fig. 6 is shown as a kind of structural schematic diagram of the image processing apparatus provided by the invention in an embodiment.
Component label instructions
601 processing units
602 determination units
603 adjustment units
S101~S105 step
Specific embodiment
Illustrate embodiments of the present invention below by way of specific specific example, those skilled in the art can be by this specification
Other advantages and efficacy of the present invention can be easily understood for disclosed content.The present invention can also pass through in addition different specific realities
The mode of applying is embodied or practiced, the various details in this specification can also based on different viewpoints and application, without departing from
Various modifications or alterations are carried out under spirit of the invention.It should be noted that in the absence of conflict, following embodiment and implementation
Feature in example can be combined with each other.
It should be noted that illustrating the basic structure that only the invention is illustrated in a schematic way provided in following embodiment
Think, only shown in schema then with related component in the present invention rather than component count, shape and size when according to actual implementation
Draw, when actual implementation kenel, quantity and the ratio of each component can arbitrarily change for one kind, and its assembly layout kenel
It is likely more complexity.
In order to make photo effect close with the perception of human eye, needs to carry out colour cast adjustment to image, reach the effect of white balance
Fruit.But existing automatic white balance algorithm is bad to image processing effect, therefore the present invention is to improve image rectification effect,
A kind of area of skin color acquisition adjusted value using in image is proposed, and then corresponding colour cast tune is carried out to image according to adjusted value
It is whole, the method for the image after obtaining white balance.
Referring to Fig. 1, the present invention provides a kind of image processing method based on priori knowledge, comprising the following steps:
Step S101, the first area of skin color of the image of chrominance space is obtained.
Wherein, the first area of skin color of image can be the region of personage's colour of skin in the image of chrominance space.
Specifically, studies have shown that although others colour of skin of different nationalities, all ages and classes, dissimilarity seems different,
This difference is concentrated mainly in brightness, does not consider that the skin distribution of different people under the influence of brightness is consistent, and concentrate
In a lesser region.And in chrominance space, the chrominance information of the luminance information of image and image is separated, the colour of skin exists
It is gathered in chrominance space in a lesser range, therefore, the available image to chrominance space of image processing apparatus
First area of skin color.
Further, due to including at least face's colour of skin of personage, and the coloration of face's colour of skin of personage in character image
Information distribution is consistent, therefore the first area of skin color of image is the region of the character facial colour of skin in the image of chrominance space.
Further, the first area of skin color for obtaining the image of chrominance space includes: by the image of input by rgb color sky
Between be converted to the image of chrominance space.According to the parameter information in the image of chrominance space, determine in the image of chrominance space
The first area of skin color.
Specifically, the image for the input that image processing apparatus obtains is RGB (Red, Green, Blue, red, green, blue) color
The image in space.At this point, being needed since the image of rgb color space is there is no the brightness of image and chrominance information is separated
The image of rgb color space is converted to the image of chrominance space.Image processing apparatus turns by the image of rgb color space
After being changed to the image of chrominance space, the image of chrominance space can be determined according to the parameter information in the image of chrominance space
In the first area of skin color.
Further, according to the parameter information in the image of chrominance space, first in the image of chrominance space is determined
Area of skin color includes: the parameter information in the image according to chrominance space, determines coloration sky using parametrization complexion model method
Between image in the first area of skin color.
That is, image processing apparatus, which can use parametrization complexion model method, determines in the image of chrominance space the
One area of skin color.Parametrization complexion model method is to be estimated based on probability distribution by carrying out parameter after a large amount of sample statistics
Meter, and then indicate skin distribution.Wherein, simple Gauss complexion model is a kind of parameter complexion model being most widely used.Cause
This, in the present invention, image processing apparatus can use Gauss complexion model and determine first colour of skin in the image of chrominance space
Region.Its theoretical foundation is in chrominance space, and the distribution of the chromatic component of the different colours of skin reaches unanimity, similar to dimensional Gaussian
Distribution can obtain continuous data information, form a width colour of skin similarity facial image by the probability value of calculating pixel,
Its location information is denoted as two-dimensional matrix C (x, y), determines the scatter plot of the first area of skin color, as determines the first colour of skin area
Domain.
Wherein, when the image of rgb color space is converted to the image of chrominance space, according to the difference of chrominance space, turn
Change method difference.For example, chrominance space is YCbCr (Luminance, Blue-difference chroma, Red-
Difference chroma, brightness, chroma blue, red color) color space, at this point, image processing apparatus is needed RGB
The image of color space is converted to the image of YCbCr color space.
Optionally, chrominance space is YCbCr chrominance space.
It should be noted that chrominance space can also be that other include the luminance information of image and the chrominance information of image
Color space, such as YUV (Luma and Chroma, brightness and coloration) color space or YIQ (Luminance and
Chrominance, brightness and coloration) color space, the invention is not limited in this regard.
It should be noted that image processing apparatus can also use other methods according to the parameter in the image of chrominance space
Information determines the first area of skin color in the image of chrominance space.For example, utilizing a kind of adaptive threshold of non-parametric model
Partitioning algorithm determines the first area of skin color in the image of chrominance space according to the parameter information in the image of chrominance space.
The invention is not limited in this regard.
Illustratively, chrominance space is YCbCr color space.First area of skin color is the personage in the image of chrominance space
The region of face's colour of skin.At this point, image is converted to by the image of RGB after image processing apparatus gets the image of input
The image of YCbCr color space.Specific conversion method is existing calculating, and details are not described herein.Image processing apparatus is by RGB
Image be converted to the image of YCbCr color space after, in YCbCr color space, the chromatic component Cb of the different faces colour of skin
Distribution with Cr reaches unanimity, similar to dimensional gaussian distribution, as shown in Fig. 2, can be obtained by the probability value of calculating pixel
Continuous data information, forms a width colour of skin similarity facial image, and location information is denoted as two-dimensional matrix C (x, y), draws out
The CbCr scatter plot (Cr is abscissa, and Cb is ordinate) of first area of skin color, as shown in figure 3, being to determine first colour of skin
Region.
Step S102, according to the chrominance information and preset skin color range of each pixel in the first area of skin color, is determined
The confidence value of one area of skin color.
Specifically, preset skin color range is the range for the colour of skin that researcher is counted by a large amount of Skin Color Information.
Whether image processing apparatus, can be according to the first area of skin color in preset skin color range after determining the first area of skin color
In, determine the confidence value of the first area of skin color.
Optionally, when chrominance space is YCbCr chrominance space, preset skin color range includes: 133 < Cr < 177,77
< Cb < 127 and 125 < Cb+0.6Cr < 190.
Further, according to the chrominance information and preset skin color range of each pixel in the first area of skin color, is determined
The confidence value of one area of skin color includes:
According to the chrominance information and preset skin color range of each pixel in the first area of skin color, formula is utilizedCalculate the confidence value of the first area of skin color.
Wherein, n is the sum of all pixels in area of skin color, and i is the integer greater than 0, no more than n, TiIndicate ith pixel
Confidence value, T are the confidence value of the first area of skin color.
Described in example as above, face complexion has cluster property well, one that researcher is counted in YCbCr color space
Skin color range is determined as preset skin color range and is denoted as w.At this point, w is 133 < Cr < 177；77 < Cb < 127；125 < Cb+
0.6Cr < 190.According to the cluster of human body complexion, the color of the pixel in the first area of skin color determined by image processing apparatus
Angle value should meet w, and the confidence value T of the first area of skin color can be calculated according to this.At this point, image processing apparatus can be according to
The chrominance information and preset skin color range of each pixel in one area of skin color, utilize formula
Calculate the confidence value of the first area of skin color.
Step S103, when the confidence value of the first area of skin color is not more than preset threshold, according in the first area of skin color
The position of each pixel and preset skin color range adjust the position of each pixel in the first area of skin color, obtain the second area of skin color.
Wherein, some or all of in the second area of skin color pixel in the default skin color range.Second area of skin color
Confidence value is greater than preset threshold.The relative position between each pixel in second area of skin color and each pixel in the first area of skin color
Between relative position it is corresponding.
It needs, in the relative position between each pixel in the second area of skin color and the first area of skin color between each pixel
Relative position correspondence refer to, in the relative position between each pixel in the second area of skin color and the first area of skin color between each pixel
Relative position it is identical in error range, or in proportion.
Specifically, image processing apparatus is after the confidence value for calculating the first area of skin color, it can be by the first colour of skin area
The confidence value in domain is compared with preset threshold, when the confidence value of the first area of skin color is greater than preset threshold, illustrates the
Pixel majority in one area of skin color is in preset skin color range, can determine that the image of input is not necessarily to do there is no colour cast
White balance processing.And then without executing following step.
When the confidence value of the first area of skin color is not more than preset threshold, illustrate that the pixel in the first area of skin color is basic
All outside preset skin color range, there are colour casts to need to do white balance processing, so, image processing apparatus for the image of input
The position for needing to adjust each pixel in the first area of skin color adjusts each pixel in the first area of skin color into preset range
It is whole, the second area of skin color is obtained, so that the confidence value of the second area of skin color is greater than preset threshold.In order to reduce image fault,
When adjusting the position of each pixel in the first area of skin color, it should ensure that the shape of the first area of skin color does not change as far as possible.I.e.
For the relative position of each pixel does not change in the first area of skin color, so, each in the second area of skin color
The relative position of pixel is corresponding with the relative position of each pixel in the first area of skin color.
Further, when the confidence value of the first area of skin color is not more than preset threshold, according in the first area of skin color
The position of each pixel and preset skin color range adjust the position of each pixel in the first area of skin color, obtain the second area of skin color
It include: when the confidence value of the first area of skin color is no more than preset threshold, according to the first area of skin color and preset colour of skin model
It encloses, determines the adjustment vector of the first area of skin color.Execute set-up procedure；Wherein, set-up procedure includes: according to the first area of skin color
Adjustment vector, adjust the position of each pixel in the first area of skin color, and calculate the confidence level of the first area of skin color adjusted
Value；Determine whether the confidence value of the first area of skin color adjusted is greater than preset threshold.
When the confidence value of the first area of skin color after the adjustment is greater than preset threshold, by the first area of skin color adjusted
It is determined as the second area of skin color.When the confidence value of the first area of skin color after the adjustment is not more than preset threshold, adjustment is utilized
The first area of skin color afterwards updates the first area of skin color, re-executes set-up procedure.
The adjustment vector of first area of skin color be for by each pixel in the first area of skin color to preset skin color range
The vector of interior adjustment.
Further, when the confidence value of the first area of skin color is not more than preset threshold, according to the first area of skin color and
Preset skin color range determines that the adjustment vector of the first area of skin color includes: to be not more than in the confidence value of the first area of skin color
When preset threshold, according to the position of the pixel in the first area of skin color and preset skin color range, adjustment direction is determined；Root
According to adjustment direction and it is default be sized, determine the adjustment vector of the first area of skin color.
Further, when the confidence value of the first area of skin color is not more than preset threshold, according in the first area of skin color
Pixel position and preset skin color range, determine adjustment direction include: the first area of skin color confidence value not
When greater than preset threshold, the first rectangle is determined in the first area of skin color.The second square is determined in preset skin color range
Shape.The direction that the central point of first rectangle is directed toward the central point of the second rectangle is determined as adjustment direction.
Wherein, the first rectangle is that the smallest each pixel by the first area of skin color of area of the first area of skin color is surrounded
Rectangle inside.Second rectangle is to have identical breadth length ratio, each side and the corresponding parallel area of the first rectangle with the first rectangle
Maximum rectangle.
That is, image processing apparatus the confidence value for determining the first area of skin color be not more than preset threshold when,
The position of each pixel in first area of skin color can be adjusted, so that it falls in required preset skin color range
It is interior.And when adjusting the first area of skin color, the shape of the first area of skin color should not changed, as far as possible to reduce image fault.This
When, the area minimum area-encasing rectangle R of the first area of skin color, the central point of this rectangle R can be determined in the first area of skin color
For z.Determine that there is identical breadth length ratio, the corresponding parallel maximum rectangle R ' in each side with rectangle R in preset skin color range,
The central point of this rectangle R ' is z ', at this time definition vectorBy vectorDirection be determined as adjustment direction.And it adjusts
Size is pre-set, at this point, using adjustment direction as the direction of the adjustment vector of the first area of skin color, default adjustment is big
The size of the small adjustment vector as the first area of skin color, determines the adjustment vector of the first area of skin color.Determining first
After the adjustment vector of area of skin color, set-up procedure is executed.As image processing apparatus according to the adjustment vector of the first area of skin color,
The position of each pixel in the first area of skin color is adjusted, and behind the position that each pixel in the first area of skin color is played in adjustment,
The confidence value for calculating first area of skin color adjusted, by the confidence level of calculated first area of skin color adjusted
Value is compared with preset threshold.When the confidence value of the first area of skin color after the adjustment is greater than preset threshold, illustrate to adjust
The pixel majority in the first area of skin color afterwards is in preset skin color range, at this time without to first colour of skin adjusted
Region is adjusted, at this point it is possible to which the first area of skin color adjusted is determined as the second area of skin color.After the adjustment first
When the confidence value of area of skin color is not more than preset threshold, illustrate that each pixel still exists substantially in the first area of skin color adjusted
Outside preset skin color range, the position to each pixel in the first area of skin color adjusted is needed to continue to adjust.At this point, figure
Picture processing unit can re-execute set-up procedure using the first area of skin color adjusted as first area, until after adjustment
The first area of skin color confidence value be greater than preset threshold.
It is sized that can be administrator pre-set according to actual needs it should be noted that default.
Described in example as above, image processing apparatus, can be by the first skin after the confidence value for calculating the first area of skin color
The confidence value in color region is compared with preset threshold.Assuming that the confidence value of the first area of skin color is 60%, preset threshold
It is 80%.After the confidence value 60% of first area of skin color and preset threshold 80% are compared by image processing apparatus, determine
The confidence value of the first area of skin color is small out and preset threshold, needs to carry out the position of each pixel in the first area of skin color
Adjustment.At this point, image processing apparatus needs first to determine the first area of skin color according to the first area of skin color and preset skin color range
Adjustment vector.As, image processing apparatus first determines the minimum area-encasing rectangle of CbCr scatter plot.Method particularly includes: a, figure
Area minimum polygon, i.e. polygon convex hull are determined as the point of CbCr scatter plot outermost is connected by processing unit two-by-two.b,
Image processing apparatus is using the line of the two adjacent points of convex closure as a line of rectangle.C, determine that distance exists in convex closure
The farthest point of a line of rectangle obtained in step b crosses the point and is first with the rectangle obtained in stepb in parallel
The parallel lines on side obtain the Article 2 side of rectangle.D, the point on convex closure is projected to the both sides acquired, determines respective point
Two farthest points, the excessively two o'clock are parallel to each other respectively and the vertically straight line with rectangle a line and Article 2 side, and general
Two other side of above-mentioned two straight lines as rectangle.So, can be formed and enclose the first area of skin color is to surround
The peripheral rectangle of CbCr scatter plot.E, the area of this rectangle is determined.All adjacent two o'clocks of convex closure are traversed, and straight line walks again
The smallest rectangle of area is determined as the minimum area-encasing rectangle R of CbCr scatter plot by rapid b-e, the central point of this rectangle R is z, is such as schemed
Shown in 4.
Image processing apparatus needs after determining the minimum area-encasing rectangle R of CbCr scatter plot in preset skin color range
Inside determine that there is identical breadth length ratio, the corresponding parallel maximum rectangle R ' in each side with rectangle R.At this point, image procossing
Device determines there is identical breadth length ratio, each side corresponding parallel maximum rectangle R ' with rectangle R method particularly includes: (1) right
In Fig. 4, on straight line Cb=77 and in the range of abscissa is 133 177 < < Cr, appoints and take a point C0, if its coordinate is
(x0,77), x ∈ (133,177)；(2) it crosses point C0 (x0,77) and makees two orthogonal straight line m, n, the slope point of two straight lines
Not Wei minimum area-encasing rectangle R two orthogonal sides slope, be denoted as k1, k2.Wherein, k1 is right for that longer side
The slope answered, k2 are slope corresponding to that shorter side.Therefore two vertical lines for crossing C0 point can be expressed as (in order to good
Note, transverse and longitudinal coordinate axis here are changed to xy reference axis by Cr, Cb): m:y-77=k1 (x-x0)；N:y-77=k2 (x-x0).(3)
M, n is with respectively with straight line Cr=133, and (region w) defined by face complexion range has a friendship to 125 < Cb+0.6Cr < 190
Point, is denoted as C1, C2, determines line segment C0C1, the value of C0C2.Length-width ratio is denoted as 1/w in minimum area-encasing rectangle R, works as C0C1/C0C2
Value when being equal to 1/w, that is, found out in skin color range W with rectangle R with identical breadth length ratio, the corresponding parallel maximum square in each side
Shape R ', as shown in Figure 5.At this point, the central point of this rectangle R ' is z ', at this time definition vectorBy vectorDirection it is true
It is set to adjustment direction.And it is pre-set for being sized t, at this point, using adjustment direction as the adjustment of the first area of skin color to
The first area of skin color is determined using the default t that is sized as the size of the adjustment vector of the first area of skin color in the direction of amount
Vector is adjusted, as determines the adjustment vector of CbCr scatter plot.
Image processing apparatus executes set-up procedure after determining the adjustment vector of the first area of skin color.As, at image
Device is managed according to the adjustment vector of the first area of skin color, by each pixel in the first area of skin color along the first area of skin color
Adjust the vector direction translation adjustment vector magnitude of vector, the first area of skin color after being adjusted.Image processing apparatus can be with
According to the chrominance information and preset skin color range of each pixel of the first area of skin color adjusted, formula is utilized Calculate the confidence value of the first area of skin color adjusted.If adjusted
The confidence value of one area of skin color is 70%, and image processing apparatus is by the confidence value of the first area of skin color adjusted at this time
70% is compared with preset threshold 80%, determines that the confidence value 70% of the first area of skin color adjusted is less than default threshold
Value 80% as updates first at this point, the first area of skin color adjusted is determined as the first area of skin color by image processing apparatus
Area of skin color, and re-execute set-up procedure.As, image processing apparatus is according to the adjustment vector of the first area of skin color, again
Each pixel in first area of skin color is big along the vector direction translation adjustment vector of the adjustment vector of the first area of skin color
It is small, retrieve the first area of skin color adjusted.Image processing apparatus can be again according to the first area of skin color adjusted
Each pixel chrominance information and preset skin color range, utilize formulaWeight
Newly calculate the confidence value of the first area of skin color adjusted.If the confidence value of the first area of skin color after readjusting is
81%, at this time image processing apparatus by the confidence value 81% of the first area of skin color after readjustment and preset threshold 80% into
Row compares, and determines that the confidence value 81% of the first area of skin color adjusted is greater than preset threshold 80%, then can will again
First area of skin color adjusted is determined as the second area of skin color.
Step S104, according to the first area of skin color and the second area of skin color, the adjusted value of the image of chrominance space is determined.
Specifically, image processing apparatus is after determining the second area of skin color, it can be according to each in the second area of skin color
The location information of each pixel, determines the adjustment of the coloration of entire image in the location information of pixel and the first area of skin color
Value.When being converted due to the image of RGB image and chrominance space, it is related to luminance information, in order to reduce the distortion of image, because
This is also required to suitably adjust luminance information, therefore the luminance information of entire image need to be determined according to the adjusted value of chrominance information
Adjusted value.
Described in example as above, image processing apparatus, can be according in the second area of skin color after determining the second area of skin color
The position of each pixel in the position of each pixel and the first area of skin color, determines the adjusted value of the coloration of entire image.
As, Cr '=Cr+x, Cb '=Cb+y.Wherein, x is the adjust gain of Cr, as the Cr value and first of the pixel of second colour of skin
The difference of the Cr value of the pixel of the colour of skin.Y is the adjust gain of Cb, the Cb of the pixel of the Cb value and first colour of skin of the pixel of second colour of skin
The difference of value.Brightness also should be suitably adjusted after adjusting CbCr according to the conversion formula of RGB and YCbCr color space to reduce distortion
Value Y.Y '=Y+x+y can be enabled according to above-mentioned relation formula.
Step S105, the adjustment of image colour cast is carried out according to the adjusted value of the image of chrominance space, the figure after obtaining white balance
Picture.
Specifically, image processing apparatus is after the adjusted value for the image for determining chrominance space, it can be with above-mentioned adjusted value pair
Image carries out colour cast adjustment, and after having carried out colour cast adjustment, and the image of chrominance space can be converted to rgb color sky again
Between image, the image after obtaining white balance.
Further, due to the grey level histogram that Histogram Matching is by adjusting tri- channels image RGB, make its overlapping
Region increases, and the degree of image colour cast can weaken.Therefore, in order to realize the degree for reducing image colour cast as far as possible, figure is improved
As calibration result, when carrying out image conversion, histogram realization can use.It is specific as follows:
The adjustment of image colour cast is carried out according to the adjusted value of the image of chrominance space, the image after obtaining white balance includes: root
According to the adjusted value of the image of chrominance space, the adjustment gray value in the channel G in rgb color space is determined；According to the adjustment in the channel G
Gray value determines the adjustment gray value and channel B in the channel R in rgb color space using histogram matching；According to RGB color
The adjustment gray value in the channel G in color space, the adjustment gray value in the channel R and the adjustment gray value of channel B carry out image colour cast
Adjustment, the image of the rgb color space after obtaining white balance.
Specifically, Histogram Matching is exactly the histogram shape of image after predetermined processing as its name suggests, by with defined
Histogram matches the histogram of image to be treated.Compared to R, channel B, G channel luminance factor ratio is maximum, and G is to figure
The coloration of picture influences minimum, so generally adjusting the grey level histogram of R, B as matching template using the channel G grey level histogram.?
Application principle in white balance is, by R, Histogram Matching of the histogram to the channel G of channel B, tri- channels RGB after matching
Histogram overlapping region area it is maximum, according to the principle of white balance, the degree of image colour cast is minimum in this case, can be effective
Ground carries out white balance processing.
It by taking the Histogram Matching of channel B as an example, inputs channel B gray level function b_hist (i), as gray value is i's
Number of pixels.Need matched G channel gray level function g_hist (j), number of pixels when as gray value is j.Enable channel B
Gray scale aggregation function isThe channel G gray scale aggregation function isEnabling z is G (i)
Inverse function, i.e. z=G-1(t), the histogram functions after channel B matching are as follows: Z (r)=G-1(Br), according to the solution channel G gray scale
The inverse function of aggregation function G (t), then solution is brought into using the channel G gray scale aggregation function as dependent variable, channel B can be obtained to G
The matched gray scale aggregation function in channel finally switchs to gray level function b ' _ hist (i) of channel B again.After similarly being matched
The channel R gray level function r ' _ hist (i).It can by r ' _ hist (i), b ' _ hist (i) and g_hist (j) that are obtained after matching
Obtain the white balance processing image of Histogram Matching.
It should be noted that image processing apparatus can also use turning for the image of other methods progress rgb color space
It changes, such as the image of chrominance space is directly converted to the image of rgb color space, the invention is not limited in this regard.
Described in example as above, image processing apparatus, can be according to YCbCr's after the adjusted value for the image for determining YCbCr
The adjusted value of image determines the channel G gray value after adjustment: G '=1.164Y ' -0.813Cr ' -0.391Cb '+135.488.?
After determining the channel G adjusted gray value, the channel R gray value and tune after adjustment can be determined according to histogram matching
Channel B gray value after whole, obtaining it is whole after the channel G gray value, after adjustment after the channel R gray value and adjustment after channel B gray value,
It can obtain the white balance processing image of Histogram Matching, the image of the rgb color space after as obtaining white balance.
So, image processing apparatus of the present invention gets coloration using the first area of skin color of the image of chrominance space
The adjusted value of the image in space obtains white so as to carry out the adjustment of image colour cast according to the adjusted value of the image of chrominance space
Image after balance.It is the adjusted value that image is got according to the first area of skin color in image, Jin Erjin as in the present invention
The adjustment of row image colour cast, so that the present invention is adjusted image colour cast using relatively simple method, and accuracy is high, thus real
Show the reduction image color difference being simple and efficient, improves the purpose of image rectification effect.
The present invention provides a kind of image processing apparatus, as shown in Figure 6, comprising:
Processing unit 601, the first area of skin color of the image for obtaining chrominance space.
Specifically, processing unit 601, specifically for the image of input is converted to chrominance space by rgb color space
Image.According to the parameter information in the image of chrominance space, the first area of skin color in the image of chrominance space is determined.
Determination unit 602, the chrominance information of each pixel in the first area of skin color for being obtained according to processing unit 601
And preset skin color range, determine the confidence value of the first area of skin color.
Specifically, determination unit 602, specifically for according to the chrominance information of each pixel in the first area of skin color and default
Skin color range, utilize formulaCalculate first
The confidence value of area of skin color.
Wherein, n is the sum of all pixels in area of skin color, and i is the integer greater than 0, no more than n, TiIndicate ith pixel
Confidence value, T are the confidence value of the first area of skin color.
Optionally, chrominance space is YCbCr chrominance space.Preset skin color range includes: 133 < Cr <, 177,77 < Cb
< 127 and 125 < Cb+0.6Cr < 190.
Processing unit 601 is also used to the confidence value in the first determining area of skin color of determination unit 602 no more than default
When threshold value, according to the position of each pixel in the first area of skin color and preset skin color range, each picture in the first area of skin color is adjusted
The position of element, obtains the second area of skin color.
Wherein, for pixel in default skin color range, the second area of skin color is credible some or all of in the second area of skin color
Angle value is greater than preset threshold.In the relative position between each pixel in second area of skin color and the first area of skin color between each pixel
Relative position is corresponding.
At this point, determination unit 602, is also used to when the confidence value of the first area of skin color is no more than preset threshold, according to
First area of skin color and preset skin color range determine the adjustment vector of the first area of skin color.
Processing unit 601, when being not more than preset threshold specifically for the confidence value in the first area of skin color, according to determination
The first area of skin color and the preset skin color range that unit 602 determines, determine the adjustment vector of the first area of skin color.It executes
Set-up procedure；Wherein, set-up procedure includes: the adjustment vector according to the first area of skin color, adjusts each picture in the first area of skin color
The position of element, and calculate the confidence value of the first area of skin color adjusted；Determine the credible of the first area of skin color adjusted
Whether angle value is greater than preset threshold.When the confidence value of the first area of skin color after the adjustment is greater than preset threshold, after adjustment
The first area of skin color be determined as the second area of skin color.The confidence value of the first area of skin color after the adjustment is not more than default threshold
When value, the first area of skin color is updated using the first area of skin color adjusted, re-executes set-up procedure.
Further, determination unit 602 are not more than preset threshold specifically for the confidence value in the first area of skin color
When, according to the position of the pixel in the first area of skin color and preset skin color range, determine adjustment direction.According to adjustment side
To and it is default be sized, determine the adjustment vector of the first area of skin color.
Further, determination unit 602 are not more than preset threshold specifically for the confidence value in the first area of skin color
When, the first rectangle is determined in first area of skin color.The second rectangle is determined in preset skin color range.By first
The direction that the central point of rectangle is directed toward the central point of the second rectangle is determined as adjustment direction.
Wherein, the first rectangle is that the area of the first area of skin color is the smallest equal by each pixel in first area of skin color
It is enclosed in interior rectangle.Second rectangle is to have identical breadth length ratio, each side corresponding parallel with the first rectangle with the first rectangle
The maximum rectangle of area.
Determination unit 602 is also used to the second area of skin color obtained according to the first area of skin color and processing unit 601, really
Determine the adjusted value of the image of chrominance space.
The adjusted value of adjustment unit 603, the image of the chrominance space for being determined according to determination unit 602 carries out pattern colour
Adjustment partially, the image after obtaining white balance.
Specifically, adjustment unit 603, specifically for the adjusted value according to the image of chrominance space, rgb color space is determined
In the channel G adjustment gray value；Rgb color space is determined using histogram matching according to the adjustment gray value in the channel G
In the channel R adjustment gray value and channel B；According to the adjustment gray value in the channel G in rgb color space, the adjustment in the channel R
Gray value and the adjustment gray value of channel B carry out the adjustment of image colour cast, the image of the rgb color space after obtaining white balance.
So, image processing apparatus of the present invention gets coloration using the first area of skin color of the image of chrominance space
The adjusted value of the image in space obtains white so as to carry out the adjustment of image colour cast according to the adjusted value of the image of chrominance space
Image after balance.It is the adjusted value that image is got according to the first area of skin color in image, Jin Erjin as in the present invention
The adjustment of row image colour cast, so that the present invention is adjusted image colour cast using relatively simple method, and accuracy is high, thus real
Show the reduction image color difference being simple and efficient, improves the purpose of image rectification effect.
In conclusion a kind of image processing method and device based on priori knowledge provided by the invention effectively overcomes now
There is the various shortcoming in technology and has high industrial utilization value.
The above-described embodiments merely illustrate the principles and effects of the present invention, and is not intended to limit the present invention.It is any ripe
The personage for knowing this technology all without departing from the spirit and scope of the present invention, carries out modifications and changes to above-described embodiment.Cause
This, institute is complete without departing from the spirit and technical ideas disclosed in the present invention by those of ordinary skill in the art such as
At all equivalent modifications or change, should be covered by the claims of the present invention.
Claims (10)
1. a kind of image processing method based on priori knowledge, which comprises the following steps:
The first area of skin color of the image of chrominance space is obtained, the chrominance space is by the color of the luminance information of image and image
Spend the color space of information separation；
According to the chrominance information and preset skin color range of each pixel in first area of skin color, first colour of skin is determined
The confidence value in region；
When the confidence value of first area of skin color is not more than preset threshold, according to the position of each pixel in the first area of skin color
It sets and preset skin color range, adjusts the position of each pixel in first area of skin color, obtain the second area of skin color, wherein
Pixel is in the default skin color range some or all of in second area of skin color, the second area of skin color confidence level
Value is greater than preset threshold；The relative position between each pixel in second area of skin color and each picture in first area of skin color
Relative position between element is corresponding；
According to first area of skin color and second area of skin color, the adjusted value of the image of the chrominance space is determined；
The adjustment of image colour cast is carried out according to the adjusted value of the image of the chrominance space, the image after obtaining white balance.
2. the image processing method according to claim 1 based on priori knowledge, it is characterised in that: the acquisition coloration is empty
Between the first area of skin color of image include:
The image of input is converted to the image of chrominance space by rgb color space；
According to the parameter information in the image of the chrominance space, the first colour of skin area in the image of the chrominance space is determined
Domain.
3. the image processing method according to claim 1 based on priori knowledge, it is characterised in that: described according to described
The chrominance information and preset skin color range of each pixel in one area of skin color, determine the confidence value of first area of skin color
Include:
According to described
The chrominance information and preset skin color range of each pixel in one area of skin color calculate first area of skin color using formula
Confidence value；Wherein, n is the sum of all pixels in area of skin color, and i is the integer greater than 0, no more than n, and Ti indicates i-th of picture
The confidence value of element, T are the confidence value of first area of skin color.
4. the image processing method according to claim 1 based on priori knowledge, it is characterised in that: described described first
When the confidence value of area of skin color is not more than preset threshold, according to the position of each pixel in the first area of skin color and the preset colour of skin
Range adjusts the position of each pixel in first area of skin color, and obtaining the second area of skin color includes:
When the confidence value of first area of skin color is not more than preset threshold, according to first area of skin color and described pre-
If skin color range, determine the adjustment vector of first area of skin color；The adjustment vector of first area of skin color is to be used for
The vector that each pixel in first area of skin color is adjusted into the preset skin color range；
Execute set-up procedure；Wherein, the set-up procedure includes: to adjust institute according to the adjustment vector of first area of skin color
The position of each pixel in the first area of skin color is stated, and calculates the confidence value of the first area of skin color adjusted；Determine the tune
Whether the confidence value of the first area of skin color after whole is greater than preset threshold；
When the confidence value of first area of skin color adjusted is greater than preset threshold, by first colour of skin adjusted
Region is determined as the second area of skin color；
When the confidence value of first area of skin color adjusted is not more than preset threshold, described adjusted first is utilized
Area of skin color updates the first area of skin color, re-executes the set-up procedure.
5. the image processing method according to claim 4 based on priori knowledge, it is characterised in that: described described first
When the confidence value of area of skin color is not more than preset threshold, according to first area of skin color and the preset skin color range,
The adjustment vector for determining first area of skin color includes:
When the confidence value of first area of skin color is not more than preset threshold, according to the pixel in first area of skin color
The position of point and the preset skin color range, determine adjustment direction；
According to the adjustment direction and it is default be sized, determine the adjustment vector of first area of skin color.
6. the image processing method according to claim 5 based on priori knowledge, it is characterised in that: described described first
When the confidence value of area of skin color is not more than preset threshold, according to the position of the pixel in first area of skin color and described
Preset skin color range determines that adjustment direction includes:
When the confidence value of first area of skin color is not more than preset threshold, the is determined in first area of skin color
One rectangle；First rectangle is that the smallest each pixel by first area of skin color of area of the first area of skin color is wrapped
It is trapped among interior rectangle；
The second rectangle is determined in the preset skin color range；Second rectangle is with first rectangle with identical
Breadth length ratio, the maximum rectangle of each side parallel area corresponding with first rectangle；
The direction that the central point of first rectangle is directed toward the central point of second rectangle is determined as the adjustment direction.
7. the image processing method according to claim 2 based on priori knowledge, it is characterised in that: described according to the color
The adjusted value for spending the image in space carries out the adjustment of image colour cast, and the image after obtaining white balance includes:
According to the adjusted value of the image of the chrominance space, the adjustment gray value in the channel G in rgb color space is determined；
According to the adjustment gray value in the channel G, using histogram matching, the adjustment in the channel R in rgb color space is determined
Gray value and channel B；
According to the adjustment gray value in the channel G in the rgb color space, the adjustment gray value in the channel R and the adjustment ash of channel B
Angle value carries out the adjustment of image colour cast, the image of the rgb color space after obtaining white balance.
8. according to the method described in claim 2, it is characterized in that, the parameter in the image according to the chrominance space is believed
Breath, determines that the first area of skin color in the image of the chrominance space includes:
According to the parameter information in the image of the chrominance space, the chrominance space is determined using parametrization complexion model method
Image in the first area of skin color.
9. the image processing method according to claim 1-7 based on priori knowledge, which is characterized in that
The chrominance space is YCbCr chrominance space；
The preset skin color range includes: 133<；Cr<；177,77<；Cb<；127 and 125<；Cb+0.6Cr&
lt；190.
10. a kind of image processing apparatus characterized by comprising
Processing unit, the first area of skin color of the image for obtaining chrominance space, the chrominance space are by the brightness of image
Information and the color space of the chrominance information of image separation；
Determination unit, the chrominance information of each pixel in first area of skin color for being obtained according to the processing unit and
Preset skin color range determines the confidence value of first area of skin color；
The processing unit is also used to the confidence value in determining first area of skin color of the determination unit no more than pre-
If when threshold value, according to the position of each pixel in the first area of skin color and preset skin color range, adjusting first area of skin color
The position of interior each pixel, obtains the second area of skin color, wherein pixel is described some or all of in second area of skin color
In default skin color range, the second area of skin color confidence value is greater than preset threshold；Each picture in second area of skin color
Relative position between element is corresponding with the relative position between each pixel in first area of skin color；
The determination unit is also used to second colour of skin area obtained according to first area of skin color and the processing unit
Domain determines the adjusted value of the image of the chrominance space；
The adjusted value of adjustment unit, the image of the chrominance space for being determined according to the determination unit carries out image colour cast
Adjustment, the image after obtaining white balance.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610549868.6A CN107623845B (en) | 2016-07-13 | 2016-07-13 | A kind of image processing method and device based on priori knowledge |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610549868.6A CN107623845B (en) | 2016-07-13 | 2016-07-13 | A kind of image processing method and device based on priori knowledge |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107623845A CN107623845A (en) | 2018-01-23 |
| CN107623845B true CN107623845B (en) | 2019-09-06 |
Family
ID=61086882
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610549868.6A Active CN107623845B (en) | 2016-07-13 | 2016-07-13 | A kind of image processing method and device based on priori knowledge |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107623845B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110392245B (en) * | 2018-04-20 | 2021-07-16 | 瑞昱半导体股份有限公司 | White balance correction method based on skin color information and image processing device thereof |
| CN108810397B (en) * | 2018-04-23 | 2021-01-19 | 深圳和而泰数据资源与云技术有限公司 | Image color cast correction method and terminal equipment |
| CN110544257A (en) * | 2019-08-20 | 2019-12-06 | 成都品果科技有限公司 | Rapid skin color segmentation method |
| CN110533732B (en) * | 2019-09-04 | 2021-11-26 | 厦门美图之家科技有限公司 | Method and device for identifying skin color in image, electronic equipment and storage medium |
| CN113052893B (en) * | 2021-03-24 | 2022-05-06 | 中南大学 | Scatter diagram de-overlapping algorithm based on connected graph and convex hull |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20040009966A (en) * | 2002-07-26 | 2004-01-31 | 삼성전자주식회사 | Apparatus and method for correcting color |
| JP5113514B2 (en) * | 2007-12-27 | 2013-01-09 | キヤノン株式会社 | White balance control device and white balance control method |
| JP2009165036A (en) * | 2008-01-09 | 2009-07-23 | Sharp Corp | Image processor and computer program |
| CN102013005B (en) * | 2009-09-07 | 2013-04-24 | 泉州市铁通电子设备有限公司 | Local dynamic threshold color balance based human face detection method with polarized colored light |
| CN102946501B (en) * | 2011-05-26 | 2015-03-04 | 北京泰邦天地科技有限公司 | Color distortion correction method and device in imaging system or image output system |
-
2016
- 2016-07-13 CN CN201610549868.6A patent/CN107623845B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN107623845A (en) | 2018-01-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107623845B (en) | A kind of image processing method and device based on priori knowledge | |
| CN104182949B (en) | Image inking and fusing method and system based on histogram feature point registration | |
| CN103942755B (en) | Brightness of image adjusting method and device | |
| CN105915909B (en) | A kind of high dynamic range images layered compression method | |
| CN106886985B (en) | A kind of adaptive enhancement method of low-illumination image reducing colour cast | |
| CN103248793B (en) | The colour of skin optimization method of gamut conversion system and device | |
| US11277595B2 (en) | White balance method for image and terminal device | |
| CN106997584A (en) | A kind of haze weather image enchancing method | |
| CN107454345B (en) | White balancing treatment method, device and the terminal device of image | |
| CN103942769B (en) | A kind of satellite remote-sensing image fusion method | |
| CN106296620B (en) | A kind of color rendition method based on histogram translation | |
| CN106303483B (en) | A kind of image processing method and device | |
| CN110428371A (en) | Image defogging method, system, storage medium and electronic equipment based on super-pixel segmentation | |
| CN107180415A (en) | Skin landscaping treatment method and device in a kind of image | |
| CN107578383A (en) | A kind of low-light (level) image enhancement processing method | |
| CN110278425A (en) | Image enchancing method, device, equipment and storage medium | |
| CN102903081A (en) | Low-light image enhancement method based on red green blue (RGB) color model | |
| CN103839245B (en) | The Retinex colour-image reinforcing method at night of Corpus--based Method rule | |
| CN106611415A (en) | Detection method and device for skin area | |
| CN107578387A (en) | A kind of homomorphic filtering Enhancement Method based on hsv color space | |
| CN116681636B (en) | Light infrared and visible light image fusion method based on convolutional neural network | |
| CN109427041A (en) | A kind of image white balance method and system, storage medium and terminal device | |
| CN111968065A (en) | Self-adaptive enhancement method for image with uneven brightness | |
| CN108550124A (en) | A kind of illumination compensation and image enchancing method based on bionical helical | |
| CN106815602A (en) | A kind of runway FOD image detection method and devices based on multi-level features description |
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 |