JP7805142B2 - Image correction processing device and program - Google Patents

Description

The present invention relates to an image correction processing device and program.

Broadcast programs using HDR video (HDR stands for "High Dynamic Range") are produced with HDR reference white in mind to reduce variations in brightness between programs. When using the HLG system (HLG stands for "Hybrid Log Gamma"), the HDR reference white is 75% of the HLG video signal level (hereinafter sometimes referred to as "75% HLG").

The level of human facial skin is important when composing video. Patent Document 1 reports that the level of human facial skin, when taking HDR reference white into consideration, falls within the range of 45% HLG to 55% HLG. However, when the intentions of the video creator or the person and lighting conditions change, facial skin may be expressed at a level higher than 55% HLG. The HLG method adjusts brightness so that skin becomes white at 75% HLG, so skin expressed between 55% HLG and 75% HLG maintains its saturation as a skin color.

On the other hand, Non-Patent Document 1 reports that in programs produced using conventional SDR video (SDR stands for "Standard Dynamic Range"), the facial skin tone of women is higher than that of men.

Japanese Patent Application Laid-Open No. 2020-025241

Report ITU-R BT.2408-3 “Guidance for operational practices in HDR television production”, July 2019, ITU-R (Radiocommunication Sector of ITU).

In SDR video programs, the higher the level of a woman's facial skin, the more emphasis is placed on gradation expression and the lower the saturation. When attempting to express this using the saturation correction process described in Patent Document 1, many skin tones have a brightness lower than that equivalent to 75% HLG, and the saturation correction process cannot be applied, resulting in the problem of not being able to express skin tones with low saturation.

Furthermore, even if the brightness corresponding to 75% HLG, which is the inflection point of the saturation correction process, is reduced to the brightness corresponding to 55% HLG, the saturation is reduced equally at all hue angles, which poses the problem of reducing the saturation of colors other than facial skin.

One possible solution would be to lower the inflection point of the saturation correction process to a lightness equivalent to 55% HLG only for the hue angle range of skin tones. However, when switching from a skin hue angle to a non-skin hue angle, or vice versa, if the correction process formula results in a mathematically steep conversion, the gradation expression after conversion may become discontinuous. Furthermore, when limiting the skin tones range using lightness and hue angle, there is a risk that highly saturated colors that are not perceived as skin tones may also be subject to the correction process, even though they have the same hue angle.

The present invention was developed based on the above-mentioned problem recognition, and aims to provide an image correction processing device and program that can lower the saturation of only specific color areas when it is desired to increase the facial skin level in HDR program production and lower the saturation of facial skin, as in conventional SDR program production, according to production intent. It also prevents discontinuities from appearing in the gradation expression after saturation correction, and prevents colors other than the intended color from being corrected by the correction process.

[1] To solve the above problems, an image correction processing device according to one aspect of the present invention includes a saturation correction processing unit that calculates a correction factor f _cor for correcting the saturation of a pixel in an image whose color satisfies predetermined conditions, and corrects the saturation of the pixel based on the correction factor f _cor . The saturation correction processing unit determines the correction factor f _cor , which continuously changes in response to changes in the lightness and the hue angle, based on at least the lightness and hue angle of the pixel, and corrects the saturation of the pixel by multiplying the determined correction factor f _cor by the saturation of the _pixel .
With this configuration, the image correction processing device can correct the saturation of pixels in an image by multiplying the saturation of the pixels by the correction factor _fcor . The correction factor _fcor changes continuously in response to changes in lightness and also changes continuously in response to changes in hue angle. Therefore, no discontinuous changes in gradation occur in the corrected image.

[2] Furthermore, one aspect of the present invention is the above-mentioned image correction processing device, further comprising: a color space conversion processing unit that converts an image included in an input high dynamic range (HDR) video into a signal in a CIELAB color space corresponding to HDR and passes the signal to the saturation correction processing unit; and a color space reconversion processing unit that reconverts the signal in the CIELAB color space corresponding to HDR after processing by the saturation correction processing unit into HDR video and outputs the HDR video, and the saturation correction processing unit corrects the saturation of the pixel based on the signal in the CIELAB color space corresponding to HDR passed from the color space conversion processing unit.
Here, the color space conversion processing unit converts the image included in the CIELAB color space HDR video into a signal in the CIELAB color space in which 75% HLG scene luminance corresponds to diffuse white (lightness 100). The CIELAB color space in which 75% HLG scene luminance corresponds to diffuse white (lightness 100) is referred to as the "CIELAB color space compatible with HDR."
The saturation correction processing unit also corrects the saturation of pixels based on the signal in the CIELAB color space corresponding to the HDR, which is passed from the color space conversion processing unit.
As mentioned above, "HDR" is an abbreviation for "High Dynamic Range."

[3] Furthermore, in one aspect of the present invention, in the image correction processing device described above, the correction factor f _cor has the effect of reducing the saturation of the pixel, and the saturation correction processing unit reduces the degree of reduction in saturation caused by the correction factor f _cor as the pre-correction saturation C ^* _ab of the pixel becomes higher.

[4] Also, in one aspect of the present invention, in the above-mentioned image correction processing device, the saturation correction processing unit calculates a reduction function value C ^* _func that represents the degree of reduction in saturation due to the correction factor _fcor , according to the pre-correction saturation C ^* _ab of the pixel, and the reduction function value C ^* _func monotonically decreases as the pre-correction saturation C ^* _ab of the pixel increases, and when the pre-correction saturation C ^* _ab of the pixel is 50 or less, the reduction function value C ^* _func is 0.95 or more, and when the pre-correction saturation C ^* _ab of the pixel is 150 or more, the reduction function value C ^* _func is 0.02 or less, and the saturation correction processing unit obtains the degree of reduction in saturation due to the correction factor _fcor by multiplying a predetermined value determined based on the lightness and hue angle of the pixel by the reduction function value C ^* _func .

[5] In another aspect of the present invention, in the image correction processing device described above, the saturation correction processing unit calculates the correction factor f _cor using a predetermined formula. However, if the result of calculation using the formula is f _cor <0, f _cor is forcibly set to 0. The values appearing in the formula have the following meanings: L ^* _max is the maximum lightness; L ^* is the lightness of the pixel (0≦L ^* ≦L ^* _max ); L ^* _f is a predetermined inflection point value for lightness that is determined to change continuously in response to changes in the hue angle h _ab of the pixel; h ₁ and h ₄ are parameters appropriately determined in response to the region of the hue angle h _ab to be corrected (where h ₁ ≦h ₄ ); σ _f is a saturation correction parameter that is determined to change continuously in response to changes in the lightness of the pixel; and C ^* _func is a decreasing function value.

[6] Another aspect of the present invention is a program for causing a computer to _function as an image correction processing device, which includes a saturation correction processing unit that calculates a correction factor f _cor for correcting the saturation of a pixel of a color that meets predetermined conditions within an image, and corrects the saturation of the pixel based on the correction factor f _cor , wherein the saturation correction processing unit determines, based on at least the lightness and hue angle of the pixel, the correction factor f _cor that continuously changes in response to changes in the lightness and that continuously changes in response to changes in the hue angle, and corrects the saturation of the pixel by multiplying the saturation of the pixel by the determined correction factor f _cor .

The image correction processing device of the present invention makes it possible to correct saturation only in areas that meet specific conditions regarding hue and brightness.

1 is a block diagram showing a schematic functional configuration of an image correction processing device according to an embodiment of the present invention. 10 is an example of a graph showing the relationship between the hue angle h _ab and the lightness function value L ^* _f in the processing of the image correction processing device according to the embodiment. 10 is an example of a graph showing the relationship between the hue angle h _ab and the saturation correction parameter function value σ _f in the processing of the image correction processing device according to the embodiment. 10 is an example of a graph showing the relationship between lightness L ^* and saturation correction parameter _σf in processing by the image correction processing device according to the embodiment. 10 is a graph showing an example of the relationship between input and output of a decreasing function that has the effect of making the effect of skin color correction processing less effective as saturation increases in processing by the image correction processing device according to the embodiment. 10 is a flowchart showing a processing procedure of the image correction processing device according to the embodiment. FIG. 2 is a block diagram showing an example of the internal configuration of the image correction processing device according to the embodiment.

Next, one embodiment of the present invention will be described with reference to the drawings. The image correction processing device 1 of this embodiment corrects an input HDR (high dynamic range) video signal. Specifically, when the skin level of a human face is expressed highly in the HDR video signal, the image correction processing device 1 corrects the skin color in the image by reducing saturation and preserving the gradation. Note that in this embodiment, "skin color" refers to a color also known as pale orange, a name for a human skin color commonly seen in Japan and other Far Eastern regions. The video processed by the image correction processing device 1 is a time series of frame images. The video may or may not be accompanied by audio. When the image correction processing device 1 corrects the video signal, it performs correction on each individual frame image.

The image correction processing device 1 of this embodiment performs the following overall processing. That is, the image correction processing device 1 inputs an HDR video signal. The image correction processing device 1 then sequentially converts the input HDR video signal into a linear signal and a signal in the CIELAB color space (CIE 1976 (L*, a*, b*) color space) that supports HDR, and then corrects the saturation of only the skin color region. The image correction processing device 1 then converts the signal after saturation correction into an HDR video signal and outputs that HDR video signal. The functional configuration and processing details of the image correction processing device 1 are described below.

Figure 1 is a block diagram showing the general functional configuration of an image correction processing device according to this embodiment. As shown in the figure, the image correction processing device 1 is configured to include a color space conversion processing unit 21, a saturation correction processing unit 22, and a color space reconversion processing unit 23. Each of these functional units can be implemented using electronic circuits. Input signals to the image correction processing device 1 and output signals from the image correction processing device 1 may be electrical signals. At least a portion of the image correction processing device 1 can also be implemented using a computer and a program. Also, at least a portion of the image correction processing device 1 can be implemented using dedicated circuits (hardware). Each functional unit has storage means as needed. The storage means is implemented using, for example, semiconductor memory. Furthermore, non-volatile storage means such as a magnetic hard disk drive or solid-state drive (SSD) may also be used as needed.

The image correction processing device 1 inputs an HDR video signal (a signal standardized under the BT.2020 standard), corrects the signal, and outputs the corrected HDR video signal. Specifically, the image correction processing device 1 is configured to correct the saturation of only areas of the input HDR video signal with high facial skin levels, and then convert the corrected signal back into an HDR video signal for output (a signal standardized under the BT.2020 standard, like the input signal). The detailed functions of each component are as follows:

The color space conversion processing unit 21 converts the input HDR input video signal into a signal in the CIELAB color space that corresponds to HDR. That is, the color space conversion processing unit 21 converts the image included in the input HDR video into a signal in the CIELAB color space that corresponds to HDR, and passes the converted signal to the saturation correction processing unit 22.
The color space conversion processing unit 21 converts an image included in the CIELAB color space HDR video into a signal in the CIELAB color space in which 75% HLG scene luminance corresponds to diffuse white (lightness 100). The CIELAB color space in which 75% HLG scene luminance corresponds to diffuse white (lightness 100) is referred to as the "CIELAB color space compatible with HDR."

Specifically, the color space conversion processing unit 21 applies an EOTF function (electro-optical transfer function) conforming to the HDR standard to the HDR input video signal _E'hin = { _R'HDRin , _G'HDRin , _B'HDRin } to obtain a display luminance signal _FdHDRin = { _RHDRin , _GHDRin , _BHDRin }. Furthermore, the color space conversion processing unit 21 converts the display luminance signal _FdHDRin into tristimulus values _XHDRin , _YHDRin , _ZHDRin through matrix calculation. The color space conversion processing unit 21 then converts these tristimulus values into signals in the CIELAB color space that correspond to HDR. This processing by the color space conversion processing unit 21 calculates lightness L ^* and color coordinates a ^* and b ^* .

Note that a ^* is a value that indicates the degree of red (positive a ^* ) and green (negative a ^* ). Also, b ^* is a value that indicates the degree of yellow (positive b ^* ) and blue (negative b ^* ). Once the color coordinates a ^* and b ^* are determined, the hue angle and saturation are determined.

During this processing, the color space conversion processing unit 21 may convert to a scene luminance signal using an inverse OETF function compliant with the HDR method to obtain an RGB linear signal, and then further convert to a CIELAB color space in which 75% HLG scene luminance corresponds to diffuse white (lightness 100).

The processing performed by the color space conversion processing unit 21 itself is conventional technology and is equivalent to the processing described in Patent Document 1, for example.

The color space conversion processing unit 21 passes the CIELAB space signal corresponding to HDR obtained by the above processing to the saturation correction processing unit 22.

The saturation correction processing unit 22 calculates a correction factor f _cor for correcting the saturation of a pixel of a color that meets predetermined conditions in an image, and corrects the saturation of the pixel based on the correction factor f _cor . The saturation correction processing unit 22 determines the correction factor f _cor that continuously changes in response to changes in the lightness and also continuously changes in response to changes in the hue angle, based on at least the lightness and hue angle of the pixel. The saturation correction processing unit 22 corrects the saturation of the pixel by multiplying the _saturation of the pixel by the determined correction factor f _cor . The saturation correction processing unit 22 corrects the saturation of the pixel based on the signal in the CIELAB color space corresponding to the HDR, which is passed from the color space conversion processing unit 21.

The saturation correction processor 22 of this embodiment corrects saturation in the skin-color region. Specifically, the saturation correction processor 22 performs processing to correct the saturation of only the skin-color region exceeding 55% HLG in the HDR-compatible CIELAB space by equal lightness/equal hue conversion. When an HDR signal is converted into the HDR-compatible CIELAB color space, the skin-color region exceeding 55% HLG falls within a lightness range from L ^* _min to L ^* _max and a hue angle range from hab _,min to _hab,max . Here, lightness L ^* _min is the lightness corresponding to 55% HLG, and is L ^* _min = 63 when calculated from the display luminance and L ^* _min = 76 when calculated from the scene luminance. The lightness L ^* _max is the lightness corresponding to the upper limit of the video signal, 109% HLG. When calculated from the display luminance, L ^* _max = 224, and when calculated from the scene luminance, L ^* _max = 217. The lower limit h _ab,min and upper limit h _ab,max of the hue angle range of skin color in the CIELAB color space corresponding to HDR are h _ab,min = 35° and h _ab,max = 65°. Here, L ^* _min , L ^* _max , h _ab,min , and h _ab,max have been described as fixed values, but each of these may also be parameters that the user can freely set. Furthermore, although the upper limit of the video signal is set to 109% HLG, it may also be set to 100% HLG.

In the saturation correction process by the saturation correction processing unit 22, as shown in the following formula (1), chroma C ^* _ab and hue angle h _ab are calculated, and a correction factor f cor for reducing the saturation of skin _color may be calculated only when the hue angle h _ab is in the range of h _ab ≧ h _ab,min and h _ab ≦ h _ab,max and the lightness L ^* is L ^* > L ^* _min . However, if the calculation result according to formula (1) is f _cor < 0, f _cor is forcibly set to 0.

Furthermore, as shown in equation (1), if the above conditions are not met, the value of the correction factor f _cor is set to 1 (saturation is not corrected).

Note that in equation (1), σ is a user parameter that can be adjusted arbitrarily, and σ≧0.

When performing correction as in Equation (1) above, the hue angle h _ab changes so that the saturation drops suddenly (discontinuously) at the boundary between h _ab,min and h _ab,max . This may result in discontinuity in the gradation expression after conversion. Therefore, in this embodiment, the lightness L ^* _f at the inflection point is expressed as a function as shown in Equation (2) below, so that the lightness at the inflection point changes continuously depending on the hue angle.

Here, L ^* _ref is the lightness corresponding to 75% HLG. Furthermore, L ^* _ip is the lightness corresponding to 55% HLG. Furthermore, _h1 , _h2 , _h3 , and _h4 are hue angles that are inflection points on the hue angle side of the lightness that is the inflection point. _h1 , _h2 , _h3 , and _h4 are user parameters that can be arbitrarily adjusted by the user as long as they respectively satisfy the relationships _{h1 ≦ h ab,min} ≦ _h2 , _{h3 ≦} h _{ab ,max} ≦ h4, and h1 _{≦ h2} < h3 _{≦ h4} . In particular, a constraint may be set that limits the case to h1 < _h2 < _h3 < _h4 , excluding the case where _h1 = h2 or h3 = _h4 . As an example, h ₁ =34°, h ₂ =38°, h ₃ =62°, h ₄ =66°, etc. may be used.

FIG. 2 is a graph showing the relationship between the hue angle h _ab and the lightness function value L ^* _f for the above example (h ₁ = 34°, h ₂ = 38°, h ₃ = 62°, h ₄ = 66°). In this graph, the horizontal axis represents the hue angle h _ab in degrees (°). The range of the hue angle h _ab is 0≦h _ab < 360. The vertical axis represents the lightness L ^* _f . The lightness L ^* _f is an inflection point determined according to the hue angle h _ab . In this graph, the thick solid line represents the lightness of the inflection point according to this embodiment (Equation (2)). The thin solid line represents the lightness of the inflection point when Equation (1) is used. The dashed line represents L ^* _ref (lightness equivalent to 75% HLG). The dashed dotted line indicates L ^* _ip (lightness corresponding to 55% HLG). As shown in the figure, L ^* _ref =100 and L ^* _ip =63.

When formula (1) is used (thin solid line graph), the lightness L ^{*f of the inflection point changes discontinuously at the hue angles h ab,min and h ab,max. In contrast, when formula (2) is used (thick solid line graph), the lightness L*} _{f of the} ^inflection _point changes continuously near the hue angles h _ab,min and h _ab,max .

That is, in the thick solid line graph, when the hue angle h _ab is from 0° to h ₁ , the lightness L ^* _f at the inflection point is constant at L ^* _ref (lightness corresponding to 75% HLG). When the hue angle h _ab is from h ₁ to h ₂ (h ₁ ≦ h _ab,min ≦ h ₂ ), the lightness L ^* _f at the inflection point changes (decreases) linearly from L ^* _ref (lightness corresponding to 75% HLG) to L ^* _ip (lightness corresponding to 55% HLG). When the hue angle h _ab is from h ₂ to h ₃ , the lightness L ^* _f at the inflection point is constant at L ^* _ip (lightness corresponding to 55% HLG). When the hue angle h _ab is between _h3 and _h4 ( _h3 ≦ h _ab,max ≦ _h4 ), the lightness L ^* _f of the inflection point changes (increases) linearly from L ^* _ip (the lightness corresponding to 55% HLG) to L ^* _ref (the lightness corresponding to 75% HLG). In the region where the hue angle h _ab is greater than _h4 , the lightness L ^* _f of the inflection point is again constant at L ^* _ref (the lightness corresponding to 75% HLG).

Note that the method of linearly changing the lightness L ^* _f of the inflection point for each section (each range of hue angles h, _ab ) as shown in this graph is just one example, and it is not necessary to use such a method. However, the method of determining the lightness L ^* _f of the inflection point as shown in this graph (as shown in formula (2)) can simplify the calculation.

That is, although formula (2) was shown above as an example of a function for calculating the lightness L ^* _f of the inflection point, in general, the function for calculating the lightness L ^* _f of the inflection point is desirably as follows. That is, the function for calculating the lightness L ^* _f of the inflection point is desirably a continuous function over the entire range of the domain of the hue angle h _ab . Furthermore, in the range h ₁ ≦ h _ab ≦ h ₂ , the lightness L ^* _f of the inflection point decreases (may be monotonically decreasing) from L ^* _ref (or its vicinity) to L ^* _ip (or its vicinity). Note that the range h ₁ ≦ h _ab ≦ h ₂ includes the point h _ab = h _ab,min . Furthermore, in the range h ₂ < h _ab < h ₃ , the lightness L ^* _f of the inflection point is constant (or approximately constant) at L ^* _ip (or its vicinity). Furthermore, in the region _h3 ≦ _hab ≦ _h4 , the lightness L ^* _f of the inflection point increases (may increase monotonically) from L ^* _ip (or its vicinity) to L ^* _ref (or its vicinity). Note that the region h3 _{≦ hab} ≦ _h4 includes the point _hab = _hab,max . In other regions (i.e., _hab < _h1 and _h4 < _hab ), the lightness L ^* _f of the inflection point is constant (or nearly constant) at L ^* _ref (or its vicinity).

Next, the amount of saturation correction is determined using the above L ^* _f as the inflection point. When performing correction using equation (1), the saturation correction parameter can also be σ (where σ>0) adjusted arbitrarily by the user. However, in the case of equation (2) above, the lightness inflection point changes continuously, so it is desirable that the saturation correction parameter also change in accordance with this change. Therefore, the saturation correction parameter is also expressed as a function so that it changes in accordance with the value of the lightness L ^* _f at the inflection point. For example, the saturation correction parameter _σf is determined as shown in equation (3) below.

In formula (3), σ ₁ and σ ₂ are parameters that are set appropriately. The values of σ ₁ and σ ₂ can be adjusted arbitrarily by the user within a range that satisfies σ ₁ ≦σ _2. Note that when the value of σ _f calculated using formula (3) is σ _f < 1, σ _f = 1 may be used.

3 is a graph showing the relationship between the hue angle h _ab and the saturation correction parameter function value σ _f when σ ₁ = 1 and σ ₂ = 2, for example. In this graph, the horizontal axis represents the hue angle h _ab in degrees (°). The domain of definition is 0≦h _ab < 360. The vertical axis represents the saturation correction parameter σ _f .

3 represents the saturation correction parameter _σf according to this embodiment, calculated using equation (3). The thin solid line represents a comparison, where saturation is an example of a positive parameter ( _σf = 1.0, constant) when equation (1) is used.

The saturation correction parameter _σf shown by the thick solid line graph in Figure 3 is calculated using a linear equation for the lightness L ^* _f shown in the graph in Figure 2. That is, when the hue angle h _ab is from 0° to h ₁ , the saturation correction parameter _σf is constant at 1.0. When the hue angle h _ab is from h ₁ to h ₂ , the saturation correction parameter _σf changes linearly (increases) from 1.0 to 2.0. When the hue angle h _ab is from h ₂ to h ₃ , the saturation correction parameter _σf is constant at 2.0. When the hue angle h _ab is from h ₃ to h ₄ , the saturation correction parameter _σf changes linearly (decreases) from 2.0 to 1.0. In the region where the hue angle h _ab is greater than h ₄ , the saturation correction parameter _σf is again constant at 1.0.

4 is a graph showing an example of the relationship between lightness L ^* and the saturation correction parameter _σf . Graph G0 shown in the figure shows the saturation correction parameter when using conventional technology, and its value is constant at 1.0 regardless of lightness L ^* . Graph G1 shows the saturation correction parameter when _σ1 = 1 and _σ2 = 2 in this embodiment. In graph G1, _σf = 2.0 when lightness L ^* = L ^* _ip (= 63), _σf changes linearly from 2.0 to 1.0 when L ^* _ip < L ^* < L ^* _ref , and is constant at _σf = 1.0 when L ^* ≧L ^* _ref (= 100). Graph G2 shows the saturation correction parameter when _σ1 = 1 and _σ2 = 3 in this embodiment. In graph G2, _σf = 3.0 when lightness L ^* = L ^* _ip (= 63), _σf changes linearly from 3.0 to 1.0 when L ^* _ip < L ^* < L ^* _ref , and is constant at _σf = 1.0 when L ^* ≥ L ^* _ref (= 100). Graph G3 shows saturation correction parameters in this embodiment when _σ1 = 1 and _σ2 = 4. In graph G2, _σf = 4.0 when lightness L ^* = L ^* _ip (= 63), _σf changes linearly from 4.0 to 1.0 when L ^* _ip < L ^* < L ^* _ref , and is constant at _σf = 1.0 when L ^* ≥ L ^* _ref (= 100). It should be noted that, in cases other than the combinations of values of σ ₁ and σ ₂ shown in FIG. 4, the saturation correction parameter σ _f can be found by equation (3) (where the lower limit of σ _f is 1.0).

The saturation correction processing unit 22 further performs correction processing to maintain the saturation of highly saturated colors that are not perceived as skin colors within the hue angle range of skin colors without changing it. To achieve this, the saturation correction processing unit 22 performs calculations using a decreasing function such that the higher the saturation, the less effective the skin color correction processing becomes.

5 is a graph showing an example of the relationship between the input and output of a decreasing function that reduces the effectiveness of skin color correction processing as the saturation increases. In this graph, the horizontal axis corresponds to the saturation C ^* _ab , which is the input value to the function, and the vertical axis corresponds to the coefficient C ^* _func , which is the output value from the function.

The effect of this decrease function is as follows. That is, as shown in the figure, in the region of 0≦C ^* _ab ≦50 where skin colors exceeding 55% HLG exist, the decrease function value C ^* _func exceeds 0.95. In other words, when C ^* _ab is approximately 50 or less, skin color correction processing is effective. Furthermore, this decrease function has the characteristic that the decrease function value C ^* _func decreases sharply in the region of C ^* _ab >50. In other words, when C ^* _ab exceeds 50, skin color correction processing becomes less effective. In particular, in the region of C ^* _ab ≧150, the decrease function value C ^* _func is 0.02 or less, and is very close to 0. In other words, when C ^* _ab is 150 or more, skin color correction processing is almost ineffective.

A decreasing function that satisfies the above requirements can be realized using, for example, a sigmoid function, as shown in equation (4) below.

C ^* _cusp,min in formula (4) is the smallest saturation among the saturations that form the gamut boundary in the hue angle range _h1 ≦ _hab ≦ _h4 to which the skin color correction processing defined in formula (2) in the CIELAB space corresponding to HDR is applied. For example, when _h1 = 34° and _h4 = 66°, C ^* _cusp,min is approximately 253.3, and the hue angle at this time is 66°.

The saturation correction processing unit 22 calculates a correction factor f cor for reducing the saturation of skin color using the lightness function L ^* _f , the saturation correction parameter function σ _f , and the reduction function C ^* _func described _above . The correction factor f _cor is calculated using, for example, the following equation (5).

However, if the result of calculation using equation (5) is f _cor < 0, f _cor is forcibly set to 0. Then, the saturation correction processing unit 22 multiplies the chroma C ^* _ab by the correction factor f _cor to obtain the corrected chroma C ^* _ab,cor , as shown in equation (6) below.

C ^* _{ab, cor} = C ^* _ab x f _cor (6)

As shown in equation (5), the correction factor f _cor has the effect of reducing the saturation of a pixel. That is, the saturation correction processing unit 22 reduces the degree of reduction in saturation caused by the correction factor f _cor as the pre-correction saturation C ^* _ab of the pixel increases.

As a specific processing method, the saturation correction processing unit 22 calculates a decrease function value C ^* _func according to the pixel's pre-correction saturation C ^* _ab . That is, the decrease function is a function according to the pixel's pre-correction saturation C ^* _ab . The decrease function value C ^* _func represents the degree of decrease in saturation due to the correction factor _fcor . As already explained with reference to the example of FIG. 5, the decrease function value C ^* _func monotonically decreases with an increase in the pixel's pre-correction saturation C ^* _ab (monotonically decreasing function). As an example, this decrease function is as follows: That is, when the pixel's pre-correction saturation C ^* _ab is 50 or less, the decrease function value C ^* _func is 0.95 or more, and when the pixel's pre-correction saturation C ^* _ab is 150 or more, the decrease function value C ^* _func is 0.02 or less. That is, in this example, the reduction function value C ^* _func decreases sharply in the region where 50≦C ^* _ab ≦150. As a result, the effect of saturation correction is limited in pixels with high saturation, and the effect of saturation correction is strong in pixels with low saturation. Note that the saturation correction processing unit 22 calculates the degree of saturation reduction due to the correction factor fcor by multiplying the reduction function value C ^* _func by a predetermined value determined based on the pixel lightness and hue angle (the product of _σf and (L ^* -L ^* _f )/(L ^* _max -L ^* _{f )} in equation (5)).

As described above, the saturation correction processing unit 22 calculates the correction factor f _cor using equation (5). However, if the result of calculation using equation (5) is f _cor <0, f _cor is forcibly set to 0. The values appearing in the equations have the following meanings: L ^* _max is the maximum lightness; L ^* is the lightness of the pixel (0≦L ^* ≦L ^* _max ); L ^* _f is a predetermined inflection point value for lightness that is determined to change continuously in response to changes in the hue angle h _ab of the pixel; h ₁ and h ₄ are parameters that are appropriately determined in response to the region of the hue angle h _ab to be corrected (where h ₁ ≦h ₄ ); σ _f is a saturation correction parameter that is determined to change continuously in response to changes in the lightness of the pixel; and C ^* _func is a decreasing function value.

The saturation correction processing unit 22 calculates a corrected LAB signal from the corrected chroma C ^* _ab,cor , the lightness L ^* , and the hue angle _hab .

Note that the saturation correction processing unit 22 does not perform saturation correction on areas other than those that meet the above correction target conditions.

The saturation correction processing unit 22 passes the video signal after saturation correction processing to the color space reconversion processing unit 23.

The color space reconversion processing unit 23 converts the signal passed from the saturation correction processing unit 22 into an HDR output video signal and outputs it. In other words, the color space reconversion processing unit 23 reconverts the CIELAB color space signal corresponding to HDR after processing by the saturation correction processing unit 22 into an HDR video and outputs it.

Specifically, the color space reconversion processing unit 23 converts the received saturation-corrected signal from the HDR-compatible CIELAB color space to tristimulus values X _HDRout, Y _HDRout, and Z _HDRout . Furthermore, the color space reconversion processing unit 23 performs a matrix operation to convert the signal into a display luminance signal Fd _HDRout = {R _HDRout , G _HDRout , B _HDRout }. This processing is equivalent to the processing for obtaining an HDR output video signal E' _hout = {R' _HDRout , G' _HDRout , B' _HDRout } processed with an inverse EOTF function compliant with the HDR standard. If the conversion to a scene luminance signal has been performed in the processing of the color space conversion processing unit 21, the color space reconversion processing unit 23 converts the signal into a video signal using an OETF function that complies with the HDR standard.

Figure 6 is a flowchart showing the processing steps of the image correction processing device 1. As shown in the figure, in step S31, the color space conversion processing unit 21 converts the input HDR input video signal into a signal in the CIELAB color space that corresponds to HDR. Next, in step S32, the saturation correction processing unit 22 corrects the saturation of the CIELAB color space signal converted in step S31 for areas in the image that satisfy certain conditions. Specifically, the saturation correction processing unit 22 corrects the saturation for skin color areas that satisfy predetermined conditions. Next, in step S33, the color space reconversion processing unit 23 converts the CIELAB color space signal corrected in step S32 into an HDR output video signal for output.

Through the processing steps described above, the image correction processing device 1 can correct images. In other words, even if the facial skin level of a person in an HDR program production is higher than the video signal level assumed in conventional technology (Patent Document 1) due to production intent, gender, etc., it is possible to reproduce skin tones similar to those in conventional SDR program production. Furthermore, the processing by the image correction processing device 1 makes it possible to perform corrections only on skin-tone areas.

Figure 7 is a block diagram showing an example of the internal configuration of the image correction processing device 1 of this embodiment. The image correction processing device 1 can be realized using a computer. As shown in the figure, the computer includes a central processing unit 901, RAM 902, input/output port 903, input/output devices 904 and 905, etc., and bus 906. The computer itself can be realized using existing technology. The central processing unit 901 executes instructions contained in a program read from RAM 902, etc. In accordance with each instruction, the central processing unit 901 writes data to RAM 902, reads data from RAM 902, and performs arithmetic and logical operations. RAM 902 stores data and programs. Each element contained in RAM 902 has an address and can be accessed using the address. Note that RAM is an abbreviation for "random access memory." The input/output port 903 is a port through which the central processing unit 901 exchanges data with external input/output devices, etc. The input/output devices 904 and 905 are input/output devices. Input/output devices 904 and 905 exchange data with the central processing unit 901 via the input/output port 903. The bus 906 is a common communication path used within the computer. For example, the central processing unit 901 reads and writes data from RAM 902 via the bus 906. Also, for example, the central processing unit 901 accesses the input/output port via the bus 906.

At least some of the functions of the image correction processing device 1 can be implemented using a computer and a program. In this case, the program for implementing this function may be recorded on a computer-readable recording medium and then loaded and executed by a computer system. Note that the term "computer system" here includes hardware such as an OS and peripheral devices. Furthermore, "computer-readable recording medium" refers to portable media such as flexible disks, optical magnetic disks, ROMs, CD-ROMs, DVD-ROMs, and USB memory, as well as storage devices such as hard disks built into computer systems. In other words, a "computer-readable recording medium" may be a non-transitory computer-readable recording medium. Furthermore, "computer-readable recording medium" may also include media that temporarily and dynamically store programs, such as communication lines used when transmitting programs over networks like the Internet or telephone lines, or media that store programs for a certain period of time, such as volatile memory within the computer systems that serve as servers or clients in such cases. The program may also be designed to implement some of the functions described above, or may be capable of implementing the functions described above in combination with a program already stored in the computer system.

The above describes an embodiment, but the present invention can also be implemented in the following variations.

In the above embodiment, specific examples of the values of various parameters (L ^* _min , L ^* _max , _hab,min , _hab,max , _h1 , _h2 , _h3 , _h4 , etc.) have been described, but different values may be used as the parameter values. Furthermore, these values may be variable.

In the above embodiment, the skin color region is the region to be subjected to saturation correction, but other regions may also be subjected to saturation correction. The color conditions of the region are expressed, for example, by color coordinates a ^* , b ^* . Alternatively, the color conditions of the region may be expressed by hue angle, saturation, or brightness.

In the above embodiment, the image correction processing device 1 was configured to include all of the color space conversion processing unit 21, saturation correction processing unit 22, and color space reconversion processing unit 23. As a variation, the image correction processing device 1 may include only the saturation correction processing unit 22. In this case, it is still possible to perform the same saturation correction as in this embodiment. Note that functions corresponding to the color space conversion processing unit 21 and color space reconversion processing unit 23 may be provided in a device external to the image correction processing device 1.

The above describes in detail an embodiment of the present invention with reference to the drawings, but the specific configuration is not limited to this embodiment and includes designs that do not deviate from the gist of the present invention.

As described above, the image correction processing device 1 of this embodiment (including modified examples) makes it possible to correct saturation only in areas of an image that satisfy specific conditions. More specifically, when the level of an area containing human skin color becomes high, the image correction processing device 1 makes it possible to correct only the area containing human skin color while maintaining the saturation of areas of other colors. This makes it possible to express skin color contained in an input HDR video signal in the same way as skin color in a conventional SDR video signal.

Furthermore, according to the image correction processing device 1 of this embodiment (including the modified examples), the numerical values of the processing results are continuous at the boundary portions of the correction processing within the image, which has the effect of enabling smooth gradation expression in the corrected image.

The image correction processing device 1 of this embodiment (including variations) has the effect of (almost) maintaining the original color of pixels that have the same hue angle as the area of the color to be corrected (for example, skin color) but do not belong to the color to be corrected (i.e., their saturation is out of range) even after correction.

Next, the effects of this embodiment will be described with reference to numerical data. Tables 1, 2, and 3 below show output saturation values, which are the results of correction processing for pixels included in the input signal to the image correction processing device 1, for comparison between the conventional technology and this embodiment. In Tables 1, 2, and 3, R, G, and B are pixel values of the R (red), G (green), and B (blue) components in the RGB color space of the HDR input signal, respectively. Furthermore, L ^* is lightness, h _ab is the hue angle, and C ^* _ab is the saturation (input side). Furthermore, C ^* _ab,out is the saturation after correction processing.

Tables 1, 2, and 3 show data for hue angles of 36°, 50°, and 66°, respectively. For each of these hue angles, 15 (3 x 5) input signals are assumed, with three lightness levels (65, 90, and 120) and five saturation levels (40, 80, 120, 160, and 200). In other words, a total of 45 input signals are assumed across three phase angles.

"Prior Art 1" is the result (output saturation value) of the technique described in Patent Document 1 (when skin color correction is attempted with a saturation correction parameter σ = 2). "Prior Art 2" is the result (output saturation value) of the technique described in Patent Document 1 when the inflection point of saturation correction is changed to a lightness equivalent to 55% HLG (similarly, when skin color correction is attempted with a saturation correction parameter σ = 2). "Present Embodiment" shows the output saturation value when skin color correction, as already explained above, is attempted. Note that the parameter values for the saturation correction process according to this embodiment are _h1 = 34°, _h2 = 38°, _h3 = 62°, _h4 = 66°, _σ1 = 1, and _σ2 = 2.

As can be seen from the results shown in Tables 1 to 3, Conventional Technique 1 was unable to reduce the saturation of skin tones for signals with a brightness of 65 or 90, whereas this embodiment was able to reduce the saturation of skin tones for signals with these brightness levels. Furthermore, Conventional Technique 2 also reduced the saturation after correction when the saturation was high (when C ^* _ab was 160 or 200, for example), whereas this embodiment was able to maintain the saturation for high saturation levels even after correction. These results demonstrate the effectiveness of this embodiment.

The present invention can be used, for example, to process video that includes specific colors (HDR video). More specifically, the image correction processing device of the present invention can be used in a wide range of fields, such as imaging devices, video signal conversion devices, and video monitoring devices. However, the scope of use of the present invention is not limited to the examples given here.

1 Image correction processing device 21 Color space conversion processing unit 22 Saturation correction processing unit 23 Color space reconversion processing unit 901 Central processing unit 902 RAM
903 Input/output port 904, 905 Input/output device 906 Bus

Claims (5)

a saturation correction processing unit that calculates a correction factor f _cor for correcting the saturation of a pixel of a color that meets a predetermined condition in the image, and corrects the saturation of the pixel based on the correction factor f _cor ;
Equipped with
the saturation correction processing unit determines, based on at least the lightness and hue angle of the pixel, the correction factor f _cor that continuously changes in response to a change in the lightness and that continuously changes in response to a change in the hue angle, and corrects the saturation of the pixel by multiplying the saturation of the pixel by the determined correction factor f _{cor ;}
The correction factor f _cor has the effect of reducing the saturation of the pixel,
the saturation correction processing unit reduces the degree of reduction in saturation due to the correction factor f _cor as the pre-correction saturation C ^* _ab of the pixel increases ,
the saturation correction processing unit calculates a decrease function value C ^* _func that indicates a degree of decrease in saturation due to the correction factor _fcor , according to a saturation C ^* _ab of the pixel before correction,
the decreasing function value C ^* _func monotonically decreases with an increase in the uncorrected saturation C ^* _ab of the pixel,
When the saturation C ^* _ab of the pixel before correction is 50 or less, the decrease function value C ^* _func is 0.95 or more,
When the saturation C ^* _ab of the pixel before correction is 150 or more, the decreasing function value C ^* _func is 0.02 or less,
the saturation correction processing unit obtains the degree of reduction in saturation due to the correction factor f _cor by multiplying a predetermined value determined based on the lightness and hue angle of the pixel by the decreasing function value C ^* _func ;
Image correction processing device. The saturation correction processing unit calculates the correction factor f _cor using the following formula:
(However, if the result of calculation using the above formula is f _cor <0, f _cor =0.)
(where L ^* _max is the maximum value of lightness, L ^* is the lightness of the pixel (0≦L ^* ≦L ^* _max ), L ^* _f is a predetermined inflection point value for lightness that is determined so as to change continuously in response to changes in the hue angle h _ab of the pixel, _h1 and _h4 are parameters that are appropriately determined in response to the region of the hue angle h _ab to be corrected (where _h1 ≦ _h4 ), _σf is a saturation correction parameter that is determined so as to change continuously in response to changes in the lightness of the pixel, and C ^* _func is the decreasing function value.)
The image correction processing device according to claim 1 . A correction factor f for correcting the saturation of pixels in an image whose color meets a predetermined condition _cor is calculated, and the saturation of the pixel is calculated by the correction factor f _cor a saturation correction processing unit that performs correction based on
a color space conversion processing unit that converts an image included in an input high dynamic range (HDR) video into a signal in a CIELAB color space corresponding to HDR and passes the converted signal to the saturation correction processing unit;
a color space reconversion processing unit that reconverts the signal in the CIELAB color space corresponding to HDR after being processed by the saturation correction processing unit into an HDR image and outputs the HDR image;
Equipped with
The saturation correction processing unit calculates the correction factor f, which continuously changes in response to a change in the lightness, based on at least the lightness and the hue angle of the pixel. _cor and the correction factor f is continuously changed in response to the change in the hue angle. _cor , and the correction factor f determined for the saturation of the pixel is _cor The saturation of the pixel is corrected by multiplying
the saturation correction processing unit corrects the saturation of the pixel based on a signal in a CIELAB color space corresponding to HDR passed from the color space conversion processing unit,
The saturation correction processing unit corrects the saturation of only a skin color region exceeding 55% HLG in the CIELAB color space corresponding to HDR passed from the color space conversion processing unit by converting to equal lightness and equal hue.
Image correction processing device. The correction factor f _cor has the effect of reducing the saturation of the pixel,
The saturation correction processing unit calculates the saturation C of the pixel before correction. ^* _ab The higher the correction factor f _cor Reduce the degree of desaturation caused by
The image correction processing device according to claim 3 . A program for causing a computer to function as the image correction processing device according to any one of claims 1 to 4 .