CN101605273A - A kind of method of evaluating colour saturation quality and subsystem - Google Patents
A kind of method of evaluating colour saturation quality and subsystem Download PDFInfo
- Publication number
- CN101605273A CN101605273A CNA2009101582776A CN200910158277A CN101605273A CN 101605273 A CN101605273 A CN 101605273A CN A2009101582776 A CNA2009101582776 A CN A2009101582776A CN 200910158277 A CN200910158277 A CN 200910158277A CN 101605273 A CN101605273 A CN 101605273A
- Authority
- CN
- China
- Prior art keywords
- saturation
- samples pictures
- value
- average
- quality
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Image Analysis (AREA)
- Spectrometry And Color Measurement (AREA)
Abstract
Embodiments of the invention disclose a kind of method and subsystem of evaluating colour saturation quality, can reflect the difference between two width of cloth contrast images more accurately, improve the accuracy of color saturation grader classification.The present embodiment disclosed method comprises: gather samples pictures from the video image of waiting to test and assess; Described samples pictures is extracted the objective indicator of the average of the average of the Cb chromatic component in saturation histogram overall target, saturation average value standard deviation integrated value, the YCbCr color space and Cr chromatic component as evaluating colour saturation quality; The grader of the objective indicator of the described extraction learning training by carrying out having supervision is classified, and the subjective test and appraisal of simulation obtain the colour saturation quality classification.The technical scheme that the embodiment of the invention provides can be used for the colour saturation quality of television image is carried out objective test and appraisal.
Description
Technical field
The present invention relates to technical field of image processing, relate in particular to a kind of method and subsystem of evaluating colour saturation quality.
Background technology
Patent application (application number: 200810184542.3 at " method and system of video image quality evaluation ", open day: 2009-05-06), image quality evaluation system (Image Quality Assessment Systems, be called for short IQAS) utilize through subjective video image of testing and assessing, therefrom extract corresponding objective indicator and carry out training study, make up expert system, thereby realize quality evaluation the unknown video image to be measured.By gathering samples pictures from video image to be measured, different aspect at the subjectivity test and appraisal extracts corresponding objective indicator respectively from described samples pictures, the corresponding objective indicator of described extraction is classified by the grader that carried out supervised learning, to simulate subjective test and appraisal video image to be measured is carried out quality evaluation, can obtain the objective evaluating result of video image to be measured by objective indicator, and can simulate subjective test and appraisal and classify by extracting different corresponding objective indicators at the different aspect of subjectivity test and appraisal, the quality that can intactly reflect video image to be measured, overcome the deficiency of subjective assessment method, improve the accuracy of objective test and appraisal.
In above-mentioned patent application, for example understand in the color saturation test and appraisal, the corresponding objective indicator of extracting from every width of cloth samples pictures is: the average and the entropy of the S component of the HSV color space of samples pictures, and these four objective indicators of average of the U of YUV color space, V component.In realizing process of the present invention, the inventor puts into practice these four objective indicators of finding described extraction can not accurately reflect difference between two width of cloth contrast images, thereby has reduced the accuracy of color saturation test and appraisal classification.
Summary of the invention
The embodiment of the invention provides a kind of method and subsystem of evaluating colour saturation quality, can reflect the difference between two width of cloth contrast images more accurately, improves the accuracy of color saturation grader classification.
For achieving the above object, the embodiment of the invention is achieved by the following technical solution:
On the one hand, provide a kind of method of evaluating colour saturation quality, comprising:
From the video image of waiting to test and assess, gather samples pictures;
Described samples pictures is extracted the objective indicator of the average of the average of the Cb chromatic component in saturation histogram overall target, saturation average value standard deviation integrated value, the YCbCr color space and Cr chromatic component as evaluating colour saturation quality;
The grader of the objective indicator of the described extraction learning training by carrying out having supervision is classified, and the subjective test and appraisal of simulation obtain the colour saturation quality classification.
Wherein, have the learning training of supervision to comprise to grader:
From video image, to the samples pictures of the video image acquisition some of every class quality through the subjective test and appraisal classification of colour saturation quality;
From every width of cloth samples pictures, extract the average of the Cb chromatic component in saturation histogram overall target, saturation average value standard deviation integrated value, the YCbCr color space and these four objective indicators of average of Cr chromatic component;
One by one described these four objective indicators of every width of cloth samples pictures are imported SVMs or artificial neural net, and tell the saturation quality category that SVMs or the each input of artificial neural net should be exported;
Until input one width of cloth wait the to test and assess samples pictures of video image, SVMs or artificial neural net are through self analyzing the colour saturation quality classification of the pairing subjective test and appraisal of correct this samples pictures of output.
Particularly, described samples pictures being extracted saturation histogram overall target comprises:
Mould with Cb chromatic component in the YCbCr color space and Cr chromatic component is represented saturation, calculates the intensity value of each pixel in the described samples pictures and the saturation average of all pixels;
With the intensity value is abscissa, and the pixel number of each intensity value correspondence is an ordinate, draws the saturation histogram of this samples pictures, and utilizes Gaussian kernel that the saturation histogram of described drafting is carried out filtering;
According to filtered described saturation histogram, search and obtain pairing intensity value of maximum crest and pixel number, the left and right sides trough of this maximum crest place waveform correspondence, the pairing pixel number of second largest crest, and maximum intensity value in the described samples pictures;
Ratio with all pixel numbers in all pixel numbers and the described samples pictures between the trough of the described left and right sides calculates the histogrammic kurtosis value of described saturation;
Difference with the saturation average of all pixels in the intensity value of described maximum crest correspondence and the described samples pictures is a molecule, and intensity value maximum in the described samples pictures is a denominator, calculates the histogrammic torsion resistance of described saturation;
Deducting the pairing pixel number of described second largest crest with the pixel number of described maximum crest correspondence is molecule, pairing pixel number of described maximum crest and the pairing pixel number of described second largest crest and be denominator, calculate the histogrammic dispersiveness of described saturation;
With described kurtosis value and described torsion resistance and as molecule, described dispersiveness calculates the histogrammic overall target of described saturation as denominator.
Particularly, described samples pictures being extracted saturation average value standard deviation integrated value comprises:
Mould with Cb chromatic component in the YCbCr color space and Cr chromatic component is represented saturation, calculates saturation average, standard deviation, maximum and the minimum value of all pixels in the described samples pictures;
As molecule, the difference of the saturation maximum of all pixels and minimum value is a denominator in the described samples pictures, calculates described saturation average value standard deviation integrated value with the product of the saturation average of all pixels in the described samples pictures and standard deviation.
Preferably, the evaluating colour saturation quality method that the embodiment of the invention provides also comprises:
Gather a plurality of samples pictures from video image to be measured, each samples pictures is extracted the described objective indicator of evaluating colour saturation quality respectively and classified the corresponding score value of at every turn being classified successively by described grader; Calculate the mean value of the described corresponding score value of repeatedly classifying, as the colour saturation quality classification of described video image to be measured.
On the other hand, provide a kind of subsystem of evaluating colour saturation quality, comprising:
Color saturation objective indicator extraction unit is used for extracting the objective indicator of the average of the average of the Cb chromatic component saturation histogram overall target, saturation average value standard deviation integrated value, the YCbCr color space and Cr chromatic component as evaluating colour saturation quality from samples pictures;
The color saturation grader is used for described objective indicator with every width of cloth samples pictures as input, through self analyzing the colour saturation quality classification of the pairing subjective test and appraisal of correct this samples pictures of output.
Wherein, described color saturation objective indicator extraction unit comprises:
Saturation histogram overall target extraction module, saturation average value standard deviation integrated value extraction module, Cb chromatic component average extraction module and Cr chromatic component average extraction module.
Particularly, described saturation histogram overall target extraction module comprises:
The saturation computation submodule is used for representing saturation with the Cb chromatic component of YCbCr color space and the mould of Cr chromatic component, calculates the intensity value of each pixel in the described samples pictures and the saturation average of all pixels;
The filtering submodule is used for being abscissa with the intensity value, and the pixel number of each intensity value correspondence is an ordinate, and the saturation histogram of this samples pictures of drafting utilizes Gaussian kernel to carry out filtering;
Search submodule, be used for according to filtered described saturation histogram, search and obtain pairing intensity value of maximum crest and pixel number, the left and right sides trough of this maximum crest place waveform correspondence, the pairing pixel number of second largest crest, and maximum intensity value in the described samples pictures;
The kurtosis value calculating sub module is used for the ratio with all pixel numbers and described all pixel numbers of samples pictures between the trough of the described left and right sides, calculates the histogrammic kurtosis value of described saturation;
The torsion resistance calculating sub module, the difference that is used for the saturation average of the intensity value of described maximum crest correspondence and described all pixels of samples pictures is a molecule, intensity value maximum in the described samples pictures is a denominator, calculates the histogrammic torsion resistance of described saturation;
Dispersed calculating sub module, being used for pixel number with described maximum crest correspondence, to deduct the pairing pixel number of described second largest crest be molecule, pairing pixel number of described maximum crest and the pairing pixel number of described second largest crest and be denominator, calculate the histogrammic dispersiveness of described saturation;
The overall target calculating sub module, be used for described kurtosis value and described torsion resistance and as molecule, described dispersiveness calculates the histogrammic overall target of described saturation as denominator.
Particularly, described saturation average value standard deviation integrated value extraction module comprises:
Saturation statistics submodule is used for representing saturation with the Cb chromatic component of YCbCr color space and the mould of Cr chromatic component, adds up saturation average, standard deviation, maximum and the minimum value of all pixels in the described samples pictures;
Average value standard deviation integrated value calculating sub module, be used for product with the saturation average of described all pixels of samples pictures and standard deviation as molecule, the difference of the saturation maximum of all pixels and minimum value is a denominator in the described samples pictures, calculates described saturation average value standard deviation integrated value.
Preferably, the evaluating colour saturation quality subsystem that the embodiment of the invention provides also comprises:
The mean value computation unit, be used for a plurality of samples pictures of gathering from video image to be measured, after passing through described color saturation objective indicator extraction unit and described color saturation grader respectively, the score value of the correspondence of at every turn being classified calculates the colour saturation quality classification of their mean value as described video image to be measured.
The technical scheme that is provided by the invention described above embodiment as can be known, by treating test and appraisal video image acquisition samples pictures, from samples pictures, extract saturation histogram overall target, saturation average value standard deviation integrated value, the average of the Cb chromatic component in the YCbCr color space and the average of Cr chromatic component are as the objective indicator of evaluating colour saturation quality, can reflect the difference between two width of cloth contrast images more accurately, then the grader of the objective indicator of the described extraction learning training by carrying out having supervision is classified, thereby obtain simulating the colour saturation quality classification of subjective test and appraisal, therefore compared with prior art, can avoid the deficiency of subjective assessment method on the one hand, improve the correctness of color saturation grader classification on the other hand.
Description of drawings
In order to be illustrated more clearly in the technical scheme of the embodiment of the invention, will to do one to the accompanying drawing of required use among the embodiment below and introduce simply.
The flow chart of the evaluating colour saturation quality method that Fig. 1 provides for the embodiment of the invention;
The filtered saturation histogram that Fig. 2 provides for the embodiment of the invention;
The function unit figure of the evaluating colour saturation quality subsystem that Fig. 3 provides for the embodiment of the invention.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the invention, the technical scheme in the embodiment of the invention is clearly and completely described.
Referring to Fig. 1, the method for a kind of evaluating colour saturation quality that the embodiment of the invention provides comprises:
For example, can adopt high-end slr camera capture video image in the darkroom to carry out the collection of samples pictures.
To a video image to be measured, the subjectivity test and appraisal that need carry out comprise different aspects such as color saturation, colour cast, color gradient, color fidelity, the colour of skin, light tone registration, definition, contrast and acutance; At the different aspect of subjectivity test and appraisal, from described samples pictures, extract different corresponding objective indicators respectively.What the embodiment of the invention was carried out is evaluating colour saturation quality, and the objective indicator of extracting from samples pictures is: the average of the Cb chromatic component in saturation histogram overall target, saturation average value standard deviation integrated value, the YCbCr color space and the average of Cr chromatic component.
In this step, have the learning training of supervision to comprise to grader:
At first by subjectivity test and appraisal will be to be tested and assessed video image quality classify, for example color saturation can be divided three classes: supersaturation, good, undersaturation.
From video image, to the samples pictures of the video image acquisition some of every class quality through the subjective test and appraisal classification of colour saturation quality.For example can adopt high-end slr camera in the darkroom, to take the photo of each class video image quality respectively, and SVMs or artificial neural net be carried out training study as samples pictures.
From every width of cloth samples pictures, extract the average of the Cb chromatic component in saturation histogram overall target, saturation average value standard deviation integrated value, the YCbCr color space and these four objective indicators of average of Cr chromatic component;
One by one described these four objective indicators of every width of cloth samples pictures are imported SVMs or artificial neural net, and tell the saturation quality category that SVMs or the each input of artificial neural net should be exported;
Until input one width of cloth wait the to test and assess samples pictures of video image, SVMs or artificial neural net are through self analyzing the colour saturation quality classification of the pairing subjective test and appraisal of correct this samples pictures of output.
That is to say, use of the input of these four objective indicators at every turn as SVMs or artificial neural net, and to tell SVMs or artificial neural net to import its output for each objective indicator should be which kind of (supersaturation, good, undersaturation).SVMs or artificial neural net utilize the ability of its powerful self study, through training study repeatedly, just can carry out correct classification to each class picture.When we imported a width of cloth and wait to test and assess the objective indicator of samples pictures, SVMs or artificial neural net just can correctly draw the pairing subjective test and appraisal classification of this samples pictures by analysis, have so just finished the learning training that supervision is arranged to grader.
How extract from samples pictures on the saturation histogram overall target, the embodiment of the invention provides following steps:
(1) represents saturation with the mould of Cb chromatic component in the YCbCr color space and Cr chromatic component, calculate the intensity value of each pixel in the described samples pictures and the saturation average of all pixels.
In the YCbCr color space, Y represents luminance component, and Cb represents the chroma blue component, and Cr represents the red color component.
The intensity value Suv=sqrt of each pixel (Cb.^2+Cr.^2), wherein sqrt is an evolution, ' .^2 ' promptly all does square operation to each element in Cb matrix and the Cr matrix for a square operation.
The intensity value of all pixels in the samples pictures is asked on average, obtained the saturation average meanSuv of all pixels in the samples pictures.
(2) to being abscissa with the intensity value, the pixel number of each intensity value correspondence is an ordinate, and the saturation histogram of this samples pictures of drafting utilizes Gaussian kernel to carry out filtering.
In order to reduce the caused histogram different wave shape of the random noise and two width of cloth contrast picture nuance, can utilize Gaussian kernel that the saturation histogram is carried out filtering.For example filtered saturation histogram as shown in Figure 2.
(3) according to filtered described saturation histogram, search and obtain pairing intensity value of maximum crest and pixel number, the left and right sides trough of this maximum crest place waveform correspondence, the pairing pixel number of second largest crest, and maximum intensity value in the described samples pictures.
Search pairing intensity value maxhx of maximum crest and pixel number maxh that the saturation histogram obtains, maximum intensity value and pixel number thereof distribute in these two the value difference representative sample pictures.
Search the left and right sides trough of the maximum crest of saturation histogram place waveform correspondence, the corresponding intensity value rangeSuvhl (intensity value shown in the figure is less) of left side trough, the corresponding intensity value rangeSuvhr (intensity value shown in the figure is bigger) of the right trough.Waveform between two intensity value is represented maximum color saturation group that distributes.
Simultaneously, from the saturation histogram, search and obtain the pairing pixel number maxhr of second largest crest, and maximum intensity value maxSuv in the described samples pictures, use in order to the calculating of back.
(4) with the ratio of all pixel numbers in all pixel numbers and the described samples pictures between the trough of the described left and right sides, calculate the histogrammic kurtosis value of described saturation.
All pixel number sizeS1 in the samples pictures, left and right sides trough (rangeSuvhl, rangeSuvhr) between all pixel number sizeS2, the computing formula of calculating the histogrammic kurtosis value fengDu of saturation is:
fengDu＝sizeS2/sizeS1。
This kurtosis value maximum color saturation groups shared ratio in whole figure of having represented to distribute.
(5) difference with the saturation average of all pixels in the intensity value of described maximum crest correspondence and the described samples pictures is a molecule, and intensity value maximum in the described samples pictures is a denominator, calculates the histogrammic torsion resistance of described saturation.
The computing formula of calculating the histogrammic torsion resistance niuQuDu of saturation is:
niuQuDu＝(maxhx-meanSuv)/maxSuv。
The histogrammic torsion resistance of this saturation is the normalized value of the saturation average alternate position spike of all pixels in the intensity value of maximum crest and the samples pictures, has positive and negative.
(6) deducting the pairing pixel number of described second largest crest with the pixel number of described maximum crest correspondence is molecule, pairing pixel number of described maximum crest and the pairing pixel number of described second largest crest and be denominator, calculate the histogrammic dispersiveness of described saturation.
The computing formula of calculating the histogrammic dispersed fenSanXing of saturation is:
fenSanXing＝(maxh-maxhr)/(maxh+maxhr)。
The histogrammic dispersiveness of this saturation has been represented the contrast of first crest and second largest crest.
(7) with described kurtosis value and described torsion resistance and as molecule, described dispersiveness calculates the histogrammic overall target of described saturation as denominator.
Calculate the histogrammic kurtosis value of saturation, torsion resistance, dispersed three the computing formula of overall target sanZongHe is:
sanZongHe＝(fengDu+niuQuDu)/fenSanXing。
The histogrammic overall target of this saturation can indicate the regularity of distribution that comprises more intensity value pixel.By comparing the histogrammic overall target of saturation of two width of cloth pictures (display frame of same width of cloth image in different TVs), can analyze the difference of the color saturation display effect of this two width of cloth picture.
How extract from samples pictures on the saturation average value standard deviation integrated value, the embodiment of the invention provides following steps:
(1) still represents saturation, calculate saturation average, standard deviation, maximum and the minimum value of all pixels in the described samples pictures with the mould of Cb chromatic component in the YCbCr color space and Cr chromatic component.
As previously mentioned, in the YCbCr color space, Y represents luminance component, and Cb represents the chroma blue component, and Cr represents the red color component.
Each pixel in the samples pictures is calculated intensity value Suv=sqrt (Cb.^2+Cr.^2), and wherein sqrt is an evolution, and ' .^2 ' promptly all does square operation to each element in Cb matrix and the Cr matrix for the some square operation.
Obtain the maximum saturation value maxSuv of all pixels in the samples pictures, minimum intensity value minSuv, average meanSuv, standard deviation stdSuv.
(2) with the product of the saturation average of all pixels in the described samples pictures and standard deviation as molecule, the difference of the saturation maximum of all pixels and minimum value is a denominator in the described samples pictures, calculates described saturation average value standard deviation integrated value.
The computing formula of calculating saturation average value standard deviation integrated value meanStdSuv is:
meanStdSuv＝meanSuv*stdSuv/(maxSuv-minSuv)。
Wherein, stdSuv/ (maxSuv-minSuv) can regard the normalized value of standard deviation as.The color saturation level height of average meanSuv presentation video integral body, and the smooth degree that standard deviation stdSuv presentation video color saturation distributes, therefore the integrated value of two values can illustrate the difference that two width of cloth pictures (display frame of same width of cloth image in different TVs) color saturation is whole and distribute by consolidated statement.
In addition, because the information that Cb chromatic component in the YCbCr color space and Cr chromatic component have comprehensively comprised the color harmony color saturation, therefore their average also can be used as the color saturation The classification basis, has also selected the objective indicator of the average meanCr of the average meanCb of the Cb chromatic component in the YCbCr color space and Cr chromatic component as evaluating colour saturation quality in the embodiment of the invention.
The method of the evaluating colour saturation quality that the embodiment of the invention provides, by treating test and appraisal video image acquisition samples pictures, from samples pictures, extract saturation histogram overall target, saturation average value standard deviation integrated value, the average of the Cb chromatic component in the YCbCr color space and the average of Cr chromatic component are as the objective indicator of evaluating colour saturation quality, can reflect the difference between two width of cloth contrast images more accurately, then the grader of the objective indicator of the described extraction learning training by carrying out having supervision is classified, thereby obtain simulating the colour saturation quality classification of subjective test and appraisal, therefore compared with prior art, can avoid the deficiency of subjective assessment method on the one hand, improve the correctness of color saturation grader classification on the other hand.
A kind of optimization embodiment, the method for the evaluating colour saturation quality that the embodiment of the invention provides also comprises:
Gather a plurality of samples pictures from video image to be measured, each samples pictures is extracted the described objective indicator of evaluating colour saturation quality respectively and classified the corresponding score value of at every turn being classified successively by described grader; Calculate the mean value of the described corresponding score value of repeatedly classifying, as the colour saturation quality classification of described video image to be measured.
For example to three classification (supersaturation of color saturation, good, undersaturation), respectively corresponding score value interval, as 2 minutes＜supersaturation＜4 minute, 4 minutes are well≤5 minute, 0 minute＜undersaturation≤2 minutes, like this, import the evaluating colour saturation quality objective indicator of a samples pictures, grader is just exported the score value of the corresponding classification of this samples pictures at every turn.To a plurality of samples pictures of gathering from same video image to be measured, grader provides corresponding score value respectively to each samples pictures, gets the final colour saturation quality classification of its mean value as this video image to be measured.This optimization embodiment can improve the accuracy of evaluating colour saturation quality, and simultaneously, the quality evaluation result who obtains according to this optimization embodiment also can be used as the foundation that subjective test and appraisal personnel screen the test and appraisal sample.
Referring to Fig. 3, the embodiment of the invention also provides a kind of subsystem of evaluating colour saturation quality, comprising:
Color saturation objective indicator extraction unit 31 is used for extracting the objective indicator of the average of the average of the Cb chromatic component saturation histogram overall target, saturation average value standard deviation integrated value, the YCbCr color space and Cr chromatic component as evaluating colour saturation quality from samples pictures;
Wherein, this color saturation objective indicator extraction unit 31 can comprise:
Saturation histogram overall target extraction module 311, saturation average value standard deviation integrated value extraction module 312, Cb chromatic component average extraction module 313 and Cr chromatic component average extraction module 3 14.
Particularly, described saturation histogram overall target extraction module 311 comprises following submodule:
The saturation computation submodule is used for representing saturation with the Cb chromatic component of YCbCr color space and the mould of Cr chromatic component, calculates the intensity value of each pixel in the described samples pictures and the saturation average of all pixels;
The filtering submodule is used for being abscissa with the intensity value, and the pixel number of each intensity value correspondence is an ordinate, and the saturation histogram of this samples pictures of drafting utilizes Gaussian kernel to carry out filtering;
Search submodule, be used for according to filtered described saturation histogram, search and obtain pairing intensity value of maximum crest and pixel number, the left and right sides trough of this maximum crest place waveform correspondence, the pairing pixel number of second largest crest, and maximum intensity value in the described samples pictures;
The kurtosis value calculating sub module is used for the ratio with all pixel numbers and described all pixel numbers of samples pictures between the trough of the described left and right sides, calculates the histogrammic kurtosis value of described saturation;
The torsion resistance calculating sub module, the difference that is used for the saturation average of the intensity value of described maximum crest correspondence and described all pixels of samples pictures is a molecule, intensity value maximum in the described samples pictures is a denominator, calculates the histogrammic torsion resistance of described saturation;
Dispersed calculating sub module, being used for pixel number with described maximum crest correspondence, to deduct the pairing pixel number of described second largest crest be molecule, pairing pixel number of described maximum crest and the pairing pixel number of described second largest crest and be denominator, calculate the histogrammic dispersiveness of described saturation;
The overall target calculating sub module, be used for described kurtosis value and described torsion resistance and as molecule, described dispersiveness calculates the histogrammic overall target of described saturation as denominator.
Described saturation average value standard deviation integrated value extraction module 312 comprises following submodule:
Saturation statistics submodule is used for representing saturation with the Cb chromatic component of YCbCr color space and the mould of Cr chromatic component, adds up saturation average, standard deviation, maximum and the minimum value of all pixels in the described samples pictures;
Average value standard deviation integrated value calculating sub module, be used for product with the saturation average of described all pixels of samples pictures and standard deviation as molecule, the difference of the saturation maximum of all pixels and minimum value is a denominator in the described samples pictures, calculates described saturation average value standard deviation integrated value.
Preferably, the subsystem of the evaluating colour saturation quality that the embodiment of the invention provides also comprises:
Mean value computation unit 33, be used for a plurality of samples pictures of gathering from video image to be measured, after passing through described color saturation objective indicator extraction unit and described color saturation grader respectively, the score value of the correspondence of at every turn being classified calculates the colour saturation quality classification of their mean value as described video image to be measured.
The evaluating colour saturation quality subsystem of the embodiment of the invention can be used as a subsystem of image quality evaluation system in application number 200810184542.3 patent applications, is used for the colour saturation quality of television image is carried out objective test and appraisal.From samples pictures, extract the objective indicator of the average of the average of the Cb chromatic component in saturation histogram overall target, saturation average value standard deviation integrated value, the YCbCr color space and Cr chromatic component by color saturation objective indicator extraction unit 31, can reflect the difference between two width of cloth contrast images more accurately as evaluating colour saturation quality; By the color saturation grader with the described objective indicator of every width of cloth samples pictures as input, through self analyzing the colour saturation quality classification of the pairing subjective test and appraisal of correct this samples pictures of output, therefore compared with prior art, can avoid the deficiency of subjective assessment method on the one hand, improve the correctness of color saturation grader classification on the other hand.
Those skilled in the art can also recognize, the unit and the performing step of each example of describing in conjunction with embodiment disclosed herein, can realize with electronic hardware, computer software or the combination of the two, for the interchangeability of hardware and software clearly is described, the composition and the step of each example described prevailingly according to function in the above description.These functions still are that software mode is carried out with hardware actually, depend on the application-specific and the design constraint of technical scheme.Those skilled in the art can use distinct methods to realize described function to each specific should being used for, but this realization should not thought and exceeds scope of the present invention.
In conjunction with the method step that embodiment disclosed herein describes, the software module that can use hardware, processor to carry out, perhaps the combination of the two is implemented.Software module can place random asccess memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or the storage medium of other form arbitrarily.
Above-mentioned specific embodiment is not in order to restriction the present invention; for those skilled in the art; all under the prerequisite that does not break away from the principle of the invention, any modification of being done, be equal to replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (10)
1, a kind of method of evaluating colour saturation quality is characterized in that, comprising:
From the video image of waiting to test and assess, gather samples pictures;
Described samples pictures is extracted the objective indicator of the average of the average of the Cb chromatic component in saturation histogram overall target, saturation average value standard deviation integrated value, the YCbCr color space and Cr chromatic component as evaluating colour saturation quality;
The grader of the objective indicator of the described extraction learning training by carrying out having supervision is classified, and the subjective test and appraisal of simulation obtain the colour saturation quality classification.
According to the method for the described evaluating colour saturation quality of claim 1, it is characterized in that 2, described have the learning training of supervision to comprise to grader:
From video image, to the samples pictures of the video image acquisition some of every class quality through the subjective test and appraisal classification of colour saturation quality;
From every width of cloth samples pictures, extract the average of the Cb chromatic component in saturation histogram overall target, saturation average value standard deviation integrated value, the YCbCr color space and these four objective indicators of average of Cr chromatic component;
One by one described these four objective indicators of every width of cloth samples pictures are imported SVMs or artificial neural net, and tell the saturation quality category that SVMs or the each input of artificial neural net should be exported;
Until input one width of cloth wait the to test and assess samples pictures of video image, SVMs or artificial neural net are through self analyzing the colour saturation quality classification of the pairing subjective test and appraisal of correct this samples pictures of output.
3, according to the method for claim 1 or 2 described evaluating colour saturation qualities, it is characterized in that, described samples pictures extracted saturation histogram overall target comprise:
Mould with Cb chromatic component in the YCbCr color space and Cr chromatic component is represented saturation, calculates the intensity value of each pixel in the described samples pictures and the saturation average of all pixels;
To being abscissa with the intensity value, the pixel number of each intensity value correspondence is an ordinate, and the saturation histogram of this samples pictures of drafting utilizes Gaussian kernel to carry out filtering;
According to filtered described saturation histogram, search and obtain pairing intensity value of maximum crest and pixel number, the left and right sides trough of this maximum crest place waveform correspondence, the pairing pixel number of second largest crest, and maximum intensity value in the described samples pictures;
Ratio with all pixel numbers in all pixel numbers and the described samples pictures between the trough of the described left and right sides calculates the histogrammic kurtosis value of described saturation;
Difference with the saturation average of all pixels in the intensity value of described maximum crest correspondence and the described samples pictures is a molecule, and intensity value maximum in the described samples pictures is a denominator, calculates the histogrammic torsion resistance of described saturation;
Deducting the pairing pixel number of described second largest crest with the pixel number of described maximum crest correspondence is molecule, pairing pixel number of described maximum crest and the pairing pixel number of described second largest crest and be denominator, calculate the histogrammic dispersiveness of described saturation;
With described kurtosis value and described torsion resistance and as molecule, described dispersiveness calculates the histogrammic overall target of described saturation as denominator.
4, according to the method for claim 1 or 2 described evaluating colour saturation qualities, it is characterized in that, described samples pictures extracted saturation average value standard deviation integrated value comprise:
Mould with Cb chromatic component in the YCbCr color space and Cr chromatic component is represented saturation, calculates saturation average, standard deviation, maximum and the minimum value of all pixels in the described samples pictures;
As molecule, the difference of the saturation maximum of all pixels and minimum value is a denominator in the described samples pictures, calculates described saturation average value standard deviation integrated value with the product of the saturation average of all pixels in the described samples pictures and standard deviation.
According to the method for claim 1 or 2 described evaluating colour saturation qualities, it is characterized in that 5, described method also comprises:
Gather a plurality of samples pictures from video image to be measured, each samples pictures is extracted the described objective indicator of evaluating colour saturation quality respectively and classified the corresponding score value of at every turn being classified successively by described grader; Calculate the mean value of the described corresponding score value of repeatedly classifying, as the colour saturation quality classification of described video image to be measured.
6, a kind of subsystem of evaluating colour saturation quality is characterized in that, comprising:
Color saturation objective indicator extraction unit is used for extracting the objective indicator of the average of the average of the Cb chromatic component saturation histogram overall target, saturation average value standard deviation integrated value, the YCbCr color space and Cr chromatic component as evaluating colour saturation quality from samples pictures;
The color saturation grader is used for described objective indicator with every width of cloth samples pictures as input, through self analyzing the colour saturation quality classification of the pairing subjective test and appraisal of correct this samples pictures of output.
According to the subsystem of the described evaluating colour saturation quality of claim 6, it is characterized in that 7, described color saturation objective indicator extraction unit comprises:
Saturation histogram overall target extraction module, saturation average value standard deviation integrated value extraction module, Cb chromatic component average extraction module and Cr chromatic component average extraction module.
According to the subsystem of the described evaluating colour saturation quality of claim 7, it is characterized in that 8, described saturation histogram overall target extraction module specifically comprises:
The saturation computation submodule is used for representing saturation with the Cb chromatic component of YCbCr color space and the mould of Cr chromatic component, calculates the intensity value of each pixel in the described samples pictures and the saturation average of all pixels;
The filtering submodule is used for being abscissa with the intensity value, and the pixel number of each intensity value correspondence is an ordinate, and the saturation histogram of this samples pictures of drafting utilizes Gaussian kernel to carry out filtering;
Search submodule, be used for according to filtered described saturation histogram, search and obtain pairing intensity value of maximum crest and pixel number, the left and right sides trough of this maximum crest place waveform correspondence, the pairing pixel number of second largest crest, and maximum intensity value in the described samples pictures;
The kurtosis value calculating sub module is used for the ratio with all pixel numbers and described all pixel numbers of samples pictures between the trough of the described left and right sides, calculates the histogrammic kurtosis value of described saturation;
The torsion resistance calculating sub module, the difference that is used for the saturation average of the intensity value of described maximum crest correspondence and described all pixels of samples pictures is a molecule, intensity value maximum in the described samples pictures is a denominator, calculates the histogrammic torsion resistance of described saturation;
Dispersed calculating sub module, being used for pixel number with described maximum crest correspondence, to deduct the pairing pixel number of described second largest crest be molecule, pairing pixel number of described maximum crest and the pairing pixel number of described second largest crest and be denominator, calculate the histogrammic dispersiveness of described saturation;
The overall target calculating sub module, be used for described kurtosis value and described torsion resistance and as molecule, described dispersiveness calculates the histogrammic overall target of described saturation as denominator.
According to the subsystem of the described evaluating colour saturation quality of claim 7, it is characterized in that 9, described saturation average value standard deviation integrated value extraction module specifically comprises:
Saturation statistics submodule is used for representing saturation with the Cb chromatic component of YCbCr color space and the mould of Cr chromatic component, adds up saturation average, standard deviation, maximum and the minimum value of all pixels in the described samples pictures;
Average value standard deviation integrated value calculating sub module, be used for product with the saturation average of described all pixels of samples pictures and standard deviation as molecule, the difference of the saturation maximum of all pixels and minimum value is a denominator in the described samples pictures, calculates described saturation average value standard deviation integrated value.
According to the subsystem of the described evaluating colour saturation quality of claim 7, it is characterized in that 10, described subsystem also comprises:
The mean value computation unit, be used for a plurality of samples pictures of gathering from video image to be measured, after passing through described color saturation objective indicator extraction unit and described color saturation grader respectively, the score value of the correspondence of at every turn being classified calculates the colour saturation quality classification of their mean value as described video image to be measured.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009101582776A CN101605273B (en) | 2009-07-23 | 2009-07-23 | Method and subsystem for evaluating colour saturation quality |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009101582776A CN101605273B (en) | 2009-07-23 | 2009-07-23 | Method and subsystem for evaluating colour saturation quality |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101605273A true CN101605273A (en) | 2009-12-16 |
| CN101605273B CN101605273B (en) | 2012-03-21 |
Family
ID=41470789
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2009101582776A Expired - Fee Related CN101605273B (en) | 2009-07-23 | 2009-07-23 | Method and subsystem for evaluating colour saturation quality |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101605273B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102005046A (en) * | 2010-11-10 | 2011-04-06 | 新太科技股份有限公司 | Novel method for detecting image color cast by color saturation detecting technology |
| CN102036099A (en) * | 2010-12-30 | 2011-04-27 | 杭州华三通信技术有限公司 | Method and device for analyzing image quality |
| CN101783971B (en) * | 2010-01-26 | 2013-07-03 | 深圳市融创天下科技股份有限公司 | Method for subjective testing of video quality |
| CN104658014A (en) * | 2015-02-10 | 2015-05-27 | 重庆金山科技（集团）有限公司 | Method for detecting in-vivo chromaticity abnormality |
| CN104881853A (en) * | 2015-05-28 | 2015-09-02 | 厦门美图之家科技有限公司 | Skin color rectification method and system based on color conceptualization |
| CN105959680A (en) * | 2016-04-29 | 2016-09-21 | 北京牡丹视源电子有限责任公司 | Method and system for evaluating color revivification degree of network camera |
| CN106157337A (en) * | 2015-04-09 | 2016-11-23 | 腾讯科技（深圳）有限公司 | Picture analyzing method and the system of analysis |
| CN114035871A (en) * | 2021-10-28 | 2022-02-11 | 深圳市优聚显示技术有限公司 | Display method and system of 3D display screen based on artificial intelligence and computer equipment |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6717698B1 (en) * | 2000-02-02 | 2004-04-06 | Eastman Kodak Company | Tone scale processing based on image modulation activity |
| KR100421023B1 (en) * | 2002-02-06 | 2004-03-04 | 삼성전자주식회사 | An apparatus and method for enhancing color temperature in color image |
| KR100943274B1 (en) * | 2002-12-20 | 2010-02-23 | 삼성전자주식회사 | Color signal correction device and a method, and image processing system using the same and a method thereof |
| CN1993977A (en) * | 2004-07-27 | 2007-07-04 | 皇家飞利浦电子股份有限公司 | Maintenance of hue in a saturation-controlled color image |
-
2009
- 2009-07-23 CN CN2009101582776A patent/CN101605273B/en not_active Expired - Fee Related
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101783971B (en) * | 2010-01-26 | 2013-07-03 | 深圳市融创天下科技股份有限公司 | Method for subjective testing of video quality |
| CN102005046A (en) * | 2010-11-10 | 2011-04-06 | 新太科技股份有限公司 | Novel method for detecting image color cast by color saturation detecting technology |
| CN102036099A (en) * | 2010-12-30 | 2011-04-27 | 杭州华三通信技术有限公司 | Method and device for analyzing image quality |
| CN104658014A (en) * | 2015-02-10 | 2015-05-27 | 重庆金山科技（集团）有限公司 | Method for detecting in-vivo chromaticity abnormality |
| CN104658014B (en) * | 2015-02-10 | 2017-09-22 | 重庆金山科技（集团）有限公司 | A kind of method for detecting colorimetric abnormality in organism |
| CN106157337A (en) * | 2015-04-09 | 2016-11-23 | 腾讯科技（深圳）有限公司 | Picture analyzing method and the system of analysis |
| CN104881853A (en) * | 2015-05-28 | 2015-09-02 | 厦门美图之家科技有限公司 | Skin color rectification method and system based on color conceptualization |
| CN104881853B (en) * | 2015-05-28 | 2018-10-16 | 厦门美图之家科技有限公司 | A kind of colour of skin antidote and system based on color generalities |
| CN105959680A (en) * | 2016-04-29 | 2016-09-21 | 北京牡丹视源电子有限责任公司 | Method and system for evaluating color revivification degree of network camera |
| CN105959680B (en) * | 2016-04-29 | 2018-12-25 | 北京牡丹视源电子有限责任公司 | A kind of method and system for evaluating and testing web camera color rendition degree |
| CN114035871A (en) * | 2021-10-28 | 2022-02-11 | 深圳市优聚显示技术有限公司 | Display method and system of 3D display screen based on artificial intelligence and computer equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101605273B (en) | 2012-03-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101605273B (en) | Method and subsystem for evaluating colour saturation quality | |
| CN101426150B (en) | Video image quality evaluation method and system | |
| RU2494460C2 (en) | Device and method to process images, and software | |
| US7936377B2 (en) | Method and system for optimizing an image for improved analysis of material and illumination image features | |
| CN106596073A (en) | Method and system for detecting image quality of optical system, and testing target plate | |
| CN103198467B (en) | Image processing apparatus and image processing method | |
| CN101853504A (en) | Image quality evaluating method based on visual character and structural similarity (SSIM) | |
| CN108109145A (en) | Picture quality detection method, device, storage medium and electronic device | |
| CN111369523B (en) | Method, system, equipment and medium for detecting cell stack in microscopic image | |
| Chen et al. | An automated bacterial colony counting and classification system | |
| Mendoza et al. | A new method for edge detection in image processing using interval type-2 fuzzy logic | |
| CN114066857A (en) | Infrared image quality evaluation method and device, electronic equipment and readable storage medium | |
| JP4383352B2 (en) | Histological evaluation of nuclear polymorphism | |
| CN110706196B (en) | Clustering perception-based no-reference tone mapping image quality evaluation algorithm | |
| CN105427306A (en) | Image analysis method and apparatus for skin lustrousness | |
| Patki et al. | Cotton leaf disease detection & classification using multi SVM | |
| CN206740363U (en) | A kind of detection optical system is as the system of matter | |
| JPWO2018203514A1 (en) | Image analysis evaluation method, computer program, image analysis evaluation device | |
| CN106441820A (en) | Display screen uniformity test method and system | |
| CN105488532B (en) | A kind of license plate screening technique and device | |
| CN111083468B (en) | Short video quality evaluation method and system based on image gradient | |
| Zhang et al. | A no-reference underwater image quality evaluator via quality-aware features | |
| Li et al. | No-Reference Underwater Image Quality Assessment Based on Quality-Aware Features | |
| JPWO2018131091A1 (en) | Image processing apparatus, image processing method, and image processing program | |
| WO2015102948A1 (en) | Urine formed element classification method and apparatus |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| ASS | Succession or assignment of patent right |
Owner name: HISENSE GROUP CO., LTD. Free format text: FORMER OWNER: QINGDAO HISENSE DIGITAL MULTIMEDIA TECHNOLOGY, THE STATE KEY LABORATORY OF INFORMATION SECURITY Effective date: 20120913 |
|
| C41 | Transfer of patent application or patent right or utility model | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20120913 Address after: 266100 Zhuzhou Road, Laoshan District, Shandong, No. 151, No. Patentee after: Hisense Group Co., Ltd. Address before: 266100 Zhuzhou Road, Laoshan District, Shandong, No. 151, No. Patentee before: Qingdao Hisense Digital Multimedia Technology, The State Key Laboratory Of Information Security |
|
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20120321 Termination date: 20200723 |