JP2014030083A - Image processor, image display device and image processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve such a problem that color space cannot be correctly converted when a color difference signal is saturated.SOLUTION: An image processor has a saturation determination part 1, a correlation analysis part 2, a color difference signal correction part 3, and a color space conversion part 4. The saturation determination part 1 outputs a result obtained by determining whether a color difference component is saturated about each pixel of an input image composed of a luminance component and a color difference component as a saturation determination result. The correlation analysis part 2 discriminates a correlation between the luminance signal and the color difference signal about each pixel of the input image, and outputs the correlation as a correlation discrimination signal. The color signal difference signal correction part 3 outputs a color difference component correction image obtained by correcting the color difference component by using the correlation between the luminance component and the color difference component on the basis of the correlation discrimination signal about a pixel showing that the color difference component is saturated by the saturation discrimination result among pixels constituting the input image. The color space conversion part 4 converts color space of the color difference component correction image.

Description

  The present invention relates to an image processing apparatus, an image processing method, and an image display apparatus that display an image in a color gamut different from the color gamut of an input image.

  A conventional display device includes a brightness adjustment unit as an image processing device when an image is displayed in a color gamut different from the color gamut of a broadcast signal or the like. When the boundary between the object color and the emission color included in the input video signal is the target line and the boundary for determining the object color and the emission color included in the output video signal is the determination line, the brightness adjustment unit When the luminance is lower than the discrimination line, the luminance of the pixel is adjusted to be lower than the target line based on the target line and the discrimination line. As described above, when an input video signal corresponding to a predetermined color space is mapped to a color space different from the predetermined color space, the conventional display device adjusts the luminance (for example, Patent Document 1).

JP 2011-124708 A

  However, the prior art as described above is a method of comparing the luminance of a pixel with a discrimination line. Since this method does not determine whether the signal levels of the two color difference signals are saturated, the input color difference signal is saturated when displaying an image in a color gamut different from the color gamut of the broadcast signal, etc. The color space could not be converted correctly.

  The present invention has been made to solve the above-described problems, and an object thereof is to correctly convert a color space even if a color difference component is saturated in an input image.

  The image processing apparatus of the present invention includes a saturation determination unit that outputs a result of determining whether or not the color difference component is saturated for each pixel of the input image composed of a luminance component and a color difference component, and a saturation determination result of the input image. For each pixel, the correlation between the luminance signal and the color difference signal is determined and output as a correlation determination signal; among the pixels constituting the input image, the color difference component is saturated by the saturation determination result A color difference signal correction unit that outputs a color difference component corrected image in which the color difference component is corrected using the correlation between the luminance component and the color difference component based on the correlation determination signal, and the color difference component And a color space conversion unit that converts the color space of the corrected image.

  According to the present invention, even when the color difference signal is saturated in the input image, the color space can be converted by correcting the saturated color difference signal.

It is a block diagram which shows an example of a structure of the image processing apparatus by Embodiment 1 of this invention. It is a table | surface which shows the relationship between a luminance component and a color difference component. 3 is a flowchart illustrating an image processing method according to the first embodiment. It is a figure showing the relationship of the area | region U (p, q) in the vicinity of the input image D0, coordinates (p, q), and (p, q).

  Hereinafter, an image processing apparatus according to the present invention will be specifically described with reference to the drawings showing embodiments thereof.

Embodiment 1 FIG.
FIG. 1 is a block diagram showing a configuration of an image processing apparatus according to Embodiment 1 of the present invention. FIG. 2 is a table showing the correlation between the luminance signal and the color difference signal. FIG. 3 is a diagram illustrating a processing procedure of the image processing method according to the first embodiment of the present invention.

  In FIG. 1, the image processing apparatus 10 includes a saturation determination unit 1, a correlation analysis unit 2, and a color difference signal correction unit 3. Further, the image processing apparatus 10 can include a color space conversion unit 4. The input image D0 is a YCbCr format signal. The saturation determination unit 1 inputs an input image D0. The saturation determination unit 1 determines whether or not the color difference components Cb and Cr are saturated for each pixel of the input image D0. The saturation determination unit 1 outputs the determination result as a saturation determination result D1.

  The method for determining whether or not the color difference component is saturated is, for example, that the color difference component is saturated if at least one of the absolute value of the color difference component Cb and the absolute value of the color difference component Cr is equal to or greater than a predetermined threshold value. It is determined that In addition to this, the method for determining whether or not the color difference component is saturated is that the color difference component is not saturated if both of the absolute values of the Cb component and the Cr component are less than a predetermined threshold value. It can also be determined. Note that, in an imaging apparatus such as a camera, it is known that the saturation value of the color difference changes depending on the luminance value. Therefore, the threshold value may be changed according to the luminance value.

  The correlation analysis unit 2 determines the correlation between the luminance signal and the color difference signal for each pixel of the input image D0 and outputs it as a correlation determination signal D2.

  Next, the operation of the correlation analysis unit 2 will be described with reference to FIG. In FIG. 2, the hue is divided into red, green, blue, cyan, magenta, yellow, white and black series. FIG. 2 shows the magnitude relationship of the R, G, and B signals for each hue. FIG. 2 shows the positive / negative relationship of signs of hue Y, Cb, and Cr signals. FIG. 2 shows the correlation between the hue Y and the Cb signal. FIG. 2 shows the correlation between the hue Y and the Cr signal. The correlation determination signal D2 is created based on the strength relationship between the hue Y and Cb signals and the strength relationship between the hue Y and Cr signals.

  For example, the formula for converting from the RGB space to the YCbCr space can be expressed by the following formula (1).

  Hereinafter, the correlation between the luminance component (Y) and the color difference components (Cb, Cr) when the color conversion is performed from the RGB space to the YCbCr space by Expression (1) will be described.

  When calculating the luminance component Y, the coefficients for weighted addition of R, G, and B are all positive. Therefore, the value of the luminance component Y is always a positive value regardless of the color. On the other hand, in the calculation of the color difference component Cb and the color difference component Cr, a positive value and a negative value are mixed in the coefficient for weighted addition of R, G, and B. For this reason, depending on the color, the values of the color difference component Cb and the color difference component Cr are positive, negative, or values close to zero. In FIG. 2, when the values of the Cb and Cr signals are close to zero, “−” is indicated. When the values of the Cb and Cr signals are positive values that are not close to zero, “positive” is indicated. When the values of the Cb and Cr signals are negative values that are not close to zero, “negative” is described.

  In the case of a red color, only the R value is large and the G and B values are small in the RGB space. In this case, the value of Cb takes a value close to zero, but Cr often takes a positive value. Considering the case where R changes to a large value, the luminance value becomes a large value, and Cb is expected to change to a large positive value while Cb remains close to zero. That is, in the red color, the value of Cb is close to zero, and the value of Cr is expected to be a large positive value. Y and Cb have no clear relationship in increase or decrease. This is because the value of Cb is close to zero regardless of the increase or decrease of Y. Y and Cr are expected to coincide with each other. This is because Cr increases as Y increases, and Cr decreases as Y decreases.

  In other words, Y and Cb have a weak correlation in reddish colors. Y and Cr have a positive correlation. That is, the increase and decrease of Y and Cr coincide.

  In the case of a green color, only the G value is large and the B value and the R value are small in the RGB space. In this case, the value of Cb takes a negative value, and Cr often takes a negative value. Considering the case where G changes to a large value, it is expected that the luminance value becomes a large value, Cb becomes a small negative value, and Cr also changes to a small negative value. That is, in the green color, Cb is a negative value and Cr is also expected to be a negative value. Y and Cb have an inverse relationship of increase / decrease. This is because Cb decreases as Y increases, and Cb increases as Y decreases. It is expected that Y and Cr have an inverse relationship of increase and decrease. This is because Cr decreases as Y increases, and Cr increases as Y decreases.

  In other words, in the green color, Y and Cb have a negative correlation in which the increase and decrease are reversed, and Y and Cr also have a negative correlation.

  Similar discussions are made for other colors. In the case of a red color, the correlation between Y and Cb is weak, and it is expected that there is a positive correlation between Y and Cr. In the case of a green color, it is expected that there is a negative correlation between Y and Cb, and there is also a negative correlation between Y and Cr. In the case of a blue color, there is a positive correlation between Y and Cb, and it is expected that Y and Cr have a weak correlation. In the case of cyan color, Y and Cb are weakly correlated, and it is expected that Y and Cr have a negative correlation. In the case of magenta colors, it is expected that there is a positive correlation between Y and Cb, and that there is also a positive correlation between Y and Cr. In the case of yellow color, there is a negative correlation between Y and Cb, and Y and Cr are expected to be weak. In the case of white or black, it is expected that Y and Cb have a weak correlation and Y and Cr have a weak correlation.

  Furthermore, in the case of a red color, Cb is close to zero, and Cr is expected to be a large positive value. For green colors, Cb is expected to be a small negative value and Cr is expected to be a small negative value. For bluish colors, Cb is a large positive value and Cr is expected to be close to zero. For cyan colors, Cb is expected to be close to zero and Cr is expected to be a small negative value. In the case of a magenta color, Cb is expected to be a large positive value, and Cr is also expected to be a large positive value. For yellow colors, Cb is a small negative value and Cr is expected to be close to zero. In the case of white or black, it is expected that Cb is close to zero and Cr is also close to zero.

  Therefore, the correlation analysis unit 2 correlates the luminance component and the color difference component according to the value of the color difference component as follows. When Cb is a large positive value, it can be determined that there is a positive correlation between the luminance component and Cb. When Cb is a small negative value, it can be determined that there is a negative correlation between the luminance component and Cb. When Cr is a large positive value, it can be determined that there is a positive correlation between the luminance component and Cr. When Cr is a small negative value, it can be determined that there is a negative correlation between the luminance component and Cr.

  The color difference signal correction unit 3 generates and outputs a color difference component corrected image D3 for each pixel of the input image D0 based on the saturation determination result D1 and the correlation determination signal D2. The color difference component corrected image D3 is an image obtained by correcting the color difference component of a pixel in which the color difference component is saturated.

  For example, when there is a positive correlation between the luminance component and the color difference component, the increase / decrease in the luminance component matches the increase / decrease in the color difference component. If there is a negative correlation between the luminance component and the color difference component, the increase / decrease in the luminance component and the increase / decrease in the color difference component are reversed.

  The corrected Cb value of the pixel whose Cb value is to be corrected is defined as Cb ′. An average value of Cb values in the vicinity of a pixel whose Cb value is to be corrected is defined as Cb_AVE. Let Y be the luminance value of the pixel whose Cb value is to be corrected, and Y_AVE be the average value of the luminance values in the vicinity of the pixel whose Cb value is to be corrected. At this time, Cb 'can be expressed by the following formula (2).

Cb ′ = Cb_AVE + Kcb · (Y−Y_AVE) (2)

  Here, Kcb is a coefficient that takes a positive value when Cb and the luminance component have a positive correlation, and Kcb is a coefficient that takes a negative value when Cb and the luminance component have a negative correlation.

  Let Cr ′ be the corrected Cr value of the pixel whose Cr value is to be corrected. An average value of Cr values in the vicinity of a pixel whose Cr value is to be corrected is defined as Cr_AVE. It is assumed that the luminance value of the pixel whose Cr value is to be corrected is Y, and the average value of the luminance values in the vicinity of the pixel whose Cr value is to be corrected is Y_AVE. At this time, Cr ′ can be expressed by the following formula (3).

Cr ′ = Cr_AVE + Kcr · (Y−Y_AVE) (3)

  Here, Kcr is a coefficient that takes a positive value when Cr and the luminance component have a positive correlation, and Kcr is a coefficient that takes a negative value when Cr and the luminance component have a negative correlation.

  The calculation of Cb_AVE is not limited to a simple average, but a weighted average may be used, and pixels in which Cb is saturated may be excluded from the calculation of the average value.

  The calculation of Cr_AVE may use not only a simple average but also a weighted average, and pixels in which Cr is saturated may be excluded from the calculation of the average value.

  In calculating Y_AVE, not only a simple average but also a weighted average may be used, and pixels in which Cb or Cr is saturated may be excluded from the calculation of the average value.

  The color space conversion unit 4 outputs an image obtained by converting the color space of the color difference component corrected image D3 as an output image D4. The color space conversion may be, for example, conversion from YCbCr space to RGB space. Note that the conversion to the RGB space is an example, and any space other than the RGB space may be used as long as it is convenient for the subsequent circuit (not shown).

  The image processing apparatus includes a saturation determination unit 1, a correlation analysis unit 2, a color difference signal correction unit 3, and a color space conversion unit 4. The saturation determination unit 1 outputs the result of determining whether or not the color difference component is saturated for each pixel of the input image D0 composed of the luminance component and the color difference component as the saturation determination result D1. The correlation analysis unit 2 determines the correlation between the luminance signal and the color difference signal for each pixel of the input image D0 and outputs it as a correlation determination signal D2. The chrominance signal correction unit 3 determines the luminance component and the chrominance component based on the correlation determination signal D2 for the pixels that constitute the input image D0 and the saturation determination result D1 indicates that the chrominance component is saturated. Is used to output a color difference component corrected image D3 in which the color difference component is corrected. The color space conversion unit 4 converts the color space of the color difference component corrected image D3. Even if the color difference component is saturated in the input image D0, the increase and decrease of the color difference component can be restored, so that the color space can be converted more correctly.

  The image processing method shown in FIG. 3 includes a saturation determination step ST1, a correlation analysis step ST2, a color difference signal correction step ST3, and a color space conversion step ST4. In the saturation determination step ST1, the saturation determination unit 1 receives the input image D0 and performs the same operation as the operation for determining whether or not the color difference components Cb and Cr are saturated for each pixel of the input image D0. In the correlation analysis step ST2, the correlation analysis unit 2 performs the same operation as the operation of determining the correlation between the luminance signal and the color difference signal for each pixel of the input image D0 and outputting it as the correlation determination signal D2. In the color difference signal correction step ST3, an image in which the color difference signal correction unit 3 corrects the color difference component of a pixel in which the color difference component is saturated based on the saturation determination result D1 and the correlation determination signal D2 for each pixel of the input image D0 is color differenced. The same operation as that output as the component correction image D3 is performed. The color space conversion step ST4 performs the same operation as the operation in which the color space conversion unit 4 outputs an image obtained by converting the color space of the color difference component corrected image D3 as the output image D4.

  The image processing apparatus shown in FIG. 1 can be used as a part of an image display apparatus such as a liquid crystal TV using a semiconductor light source such as an LED (Light Emitting Diode). Further, the image processed by the image processing method shown in FIG. 3 can be displayed on an image display device such as a liquid crystal TV using a semiconductor light source such as an LED. In an image display device using such a semiconductor light source, for example, a color gamut wider than a color gamut that can be displayed with a video signal of a high-definition television broadcast called HDTV (High Definition Television) can be realized. The color space can also be correctly converted for video signals. Such an image display apparatus has the same effect as the image table processing apparatus or the image processing method of this embodiment.

Embodiment 2. FIG.
The image processing apparatus and the image processing method according to the second embodiment can be realized with the same configuration as the image processing apparatus and the image processing method according to the first embodiment. However, the operations of the correlation analysis unit 2, the correlation analysis step ST2, the color difference signal correction unit 3, and the color difference signal correction step ST3 are different from those of the first embodiment. That is, the operation of the saturation determination unit 1, the operation of the color space conversion unit 4, the operation of the saturation determination step ST 1, and the operation of the color space conversion step ST 4 of the second embodiment are the operations of the corresponding constituent elements of the first embodiment. Is the same.

  First, the operations of the correlation analysis unit 2 and the correlation analysis step ST2 will be described. The correlation analysis unit 2 and the correlation analysis step ST2 obtain a correlation coefficient Kycb3 between Y and Cb for each pixel of the input image D0. The correlation analysis unit 2 and the correlation analysis step ST2 perform processing by regarding the Y value and Cb value of the pixels in the vicinity of each pixel of the input image D0 as random variables.

  As the correlation coefficient Kycb3, for example, a Pearson product moment correlation coefficient is used. Hereinafter, the position of each pixel constituting the input image is represented by two-dimensional coordinates (i, j), the Y value at the coordinates (i, j) is Y (i, j), and the Cb value is Cb (i, j). ). Further, the coordinates of the pixel for which the correlation coefficient Kycb3 is to be obtained are (p, q), and a set representing a pixel in the vicinity of (p, q) that is referred to for obtaining the correlation coefficient is U (p, q). When the correlation coefficient Kycb3 calculated for the pixel represented by the coordinates (p, q) is represented by Kycb3 (p, q), Kycb3 (p, q) can be represented by the following equation (4). I can do it.

  Kycb3 (p, q) takes a real value between −1 and 1. This correlation coefficient indicates that a positive correlation exists between Y and Cb when the value is close to 1, and a negative correlation exists when the value is close to −1. The average value of Y and Cb in the region U (p, q) can be expressed by the following equation (5).

  Although the average value is a simple average, the method for calculating the average value is not limited to this, and a weighted average may be used. Furthermore, when the average value is calculated, pixels in which the Cb value is saturated may be excluded.

  The correlation coefficient Kycb3 is not limited to the Pearson product moment correlation coefficient. When the value is close to 1, it indicates that there is a positive correlation between Y and Cb. It only has to indicate that there is.

  FIG. 4 is a diagram illustrating the relationship between the input image D0, the region U (p, q) in the vicinity of the coordinates (p, q), and the coordinates (p, q). In FIG. 4, the region U (p, q) is composed of three horizontal pixels and three vertical pixels. However, this is only an example, and the region U (p, q) may be a region composed of pixels in the vicinity of the coordinates (p, q).

  Further, the region U (p, q) may be composed of only pixels in which the color difference signal is not saturated among the pixels in the vicinity of the coordinates (p, q). If the region U (p, q) is configured only for pixels in which the color difference is not saturated, the correlation coefficient can be obtained more accurately.

  Alternatively, in the correlation analysis unit 2 and the correlation analysis step ST2, for each pixel of the input image D0, the area near the area is divided, and the area U (p, q) Of these, only pixels in the same region as the coordinates (p, q) may be used. Here, the region division is to divide a region in the vicinity of each pixel of the input image D0 into a pixel similar in color to the corresponding pixel (or included in the same hue) and a pixel other than that.

  The correlation coefficient Kycr3 between Y and Cr may be obtained similarly. The value of Cr at the coordinates (i, j) is represented by Cr (i, j), the position of the pixel for which the correlation coefficient Kycb3 is to be obtained is represented by coordinates (p, q), and is referred to for obtaining the correlation coefficient. A set representing pixels in the vicinity of coordinates (p, q) to be represented is represented by U (p, q), and a correlation coefficient Kycr3 calculated for the pixels represented by coordinates (p, q) is represented by Kycr3 (p, q). , Kycr3 (p, q) can be expressed by the following formula (6).

  Kycr3 (p, q) takes a real value between −1 and 1. This correlation coefficient indicates that a positive correlation exists between Y and Cb when the value is close to 1, and a negative correlation exists when the value is close to −1. The average value in the Cr region U (p, q) can be expressed by the following equation (7).

  Although the average value is a simple average, the method for calculating the average value is not limited to this, and a weighted average may be used. Furthermore, when the average value is calculated, pixels in which the Cr value is saturated may be excluded.

  The correlation coefficient Kycr3 indicates that there is a positive correlation between Y and Cb when the value is close to 1, and the correlation coefficient Kycr3 indicates that there is a negative correlation when close to −1. It does not have to be.

  Next, operations of the color difference signal correction unit 3 and the color difference signal correction step ST3 will be described. The color difference signal correction unit 3 and the color difference signal correction step ST3 correct Cb and Cr to the correlation coefficient value obtained for the corresponding pixel for each pixel of the input image D0. Here, if the correlation coefficient is a positive value, Cb and Cr may be corrected so that the increase / decrease in Cb and the increase / decrease in Cr have the same tendency as Y. If the correlation coefficient is a negative value, Cb and Cr may be corrected so that the increase and decrease of Cb and the increase and decrease of Cr have the same tendency as Y.

  If Cb is saturated for the pixel represented by coordinates (p, q), the value of Cb (p, q) is expressed as Cb ′ (p, q) represented by the following equation (8), for example. It is sufficient to correct it. Here, Kycb4 is a correction coefficient that changes in accordance with the value of the correlation coefficient Kycb3.

  When the correlation coefficient Kycb3 is a positive value, the correction coefficient Kycb4 is preferably a positive value. Alternatively, it is desirable that the correction coefficient Kycb4 does not become a negative value.

  When the correlation coefficient Kycb3 is a negative value, the correction coefficient Kycb4 is preferably a negative value. Alternatively, it is desirable that the correction coefficient Kycb4 does not become a positive value.

  Further, when the value of the correlation coefficient Kycb3 is zero, the correction coefficient Kycb4 is preferably zero. Alternatively, it is desirable that the correction coefficient Kycb4 is monotonously non-decreasing with respect to the correlation coefficient Kycb3.

  If Cr is saturated for a pixel represented by coordinates (p, q), the value of Cr (p, q) is set to, for example, Cr ′ (p, q) represented by the following equation (9). It may be corrected as follows.

  Here, Kycr4 is a correction coefficient that changes according to the value of the correlation coefficient Kycr3. When the correlation coefficient Kycr3 is a positive value, it is desirable that the correction coefficient Kycr4 also takes a positive value or does not become a negative value.

  When the correlation coefficient Kycr3 is a negative value, the correction coefficient Kycr4 is preferably a negative value. Alternatively, it is desirable that the correction coefficient Kycr4 does not become a positive value.

  When the value of the correlation coefficient Kycr3 is zero, the correction coefficient Kycr4 is preferably zero. Alternatively, it is desirable that the correction coefficient Kycr4 is monotonously non-decreasing with respect to the correlation coefficient Kycr3.

  The image processing apparatus and the image processing method according to the second embodiment corrects the color difference signal using the correlation between the luminance signal and the color difference signal. Therefore, even if the color difference component is saturated in the input image D0, the increase / decrease of the color difference component can be restored, so that the color space can be converted closer to the correct value and converted.

  The image processing apparatus according to the second embodiment can be used as a part of an image display apparatus such as a liquid crystal TV. Further, the image processed by the image processing method shown in FIG. 3 can be displayed on an image display device such as a liquid crystal TV. Such an image display apparatus has the same effect as the image table processing apparatus or image processing method of the second embodiment.

  In addition, although embodiment of this invention was described as mentioned above, this invention is not limited to these embodiment.

  1 saturation determination unit, 2 correlation analysis unit, 3 color difference signal correction unit, 4 color space conversion unit, D0 input image, D1 saturation determination result, D2 correlation discrimination signal, D3 color difference component correction image, D4 output image, ST1 saturation determination step ST2 correlation analysis step, ST3 color difference signal correction step, ST4 color space conversion step.

Claims (5)

A saturation determination unit that determines whether or not the color difference component is saturated for each pixel of the input image composed of a luminance component and a color difference component, and outputs a saturation determination result;
A correlation analysis unit that determines a correlation between the luminance signal and the color difference signal for each pixel of the input image, and outputs the correlation signal as a correlation determination signal;
Of the pixels constituting the input image, for the pixel determined to be saturated by the saturation determination result, the correlation between the luminance component and the color difference component is used based on the correlation determination signal. A color difference signal correction unit that outputs a color difference component corrected image obtained by correcting the color difference component;
An image processing apparatus comprising: a color space conversion unit that converts a color space of the color difference component corrected image. A saturation determination step of outputting, as a saturation determination result, a result of determining whether or not the color difference component is saturated for each pixel of an input image composed of a luminance component and a color difference component;
A correlation analysis step of determining the correlation between the luminance signal and the color difference signal for each pixel of the input image, and outputting the correlation determination signal;
Among the pixels constituting the input image, for the pixel indicated by the saturation determination result that the color difference component is saturated, the correlation between the luminance component and the color difference component is used based on the correlation determination signal. An image processing method comprising: a color difference signal correction step for outputting a color difference component corrected image obtained by correcting the color difference component; and a color space conversion step for converting a color space of the color difference component corrected image.   The color difference signal correcting unit corrects the color difference signal so that the increase / decrease in the luminance component and the increase / decrease in the color difference component coincide with each other when there is a positive correlation between the luminance component and the color difference component. The image processing apparatus according to claim 1.   The color difference signal correction unit corrects the color difference signal so that the increase / decrease in the luminance component and the increase / decrease in the color difference component are reversed when there is a negative correlation between the luminance component and the color difference component. The image processing apparatus according to claim 1.   An image display apparatus comprising the image processing apparatus according to claim 1.

Images