WO2023241339A1

WO2023241339A1 - Color cast correction method and apparatus, device, storage medium and program product

Info

Publication number: WO2023241339A1
Application number: PCT/CN2023/096662
Authority: WO
Inventors: 陈伟
Original assignee: 广州市百果园信息技术有限公司
Priority date: 2022-06-13
Filing date: 2023-05-26
Publication date: 2023-12-21
Also published as: CN114999363A

Abstract

The embodiments of the present application disclose a color cast correction method and apparatus, a device, a storage medium and a program product, and belong to the technical field of images. The method comprises: dynamically expanding the brightness of an image parameter of a target SDR image to obtain an expanded image parameter, the target SDR image comprising a human portrait (101); on the basis of chrominance information indicated by an image chrominance parameter in the expanded image parameter, selecting a local chrominance parameter, the local chrominance parameter being a chrominance parameter corresponding to a skin color region of the human portrait (102); performing color cast correction on the local chrominance parameter to obtain a corrected image parameter (103); and based on the corrected image parameter, generating a target HDR image corresponding to the target SDR image (104). The embodiments of the present application allow for performing color cast correction on a skin color region, thereby reducing the problem of an increase in brightness causing higher saturation of the skin color region, thus improving visual effect.

Description

Color shift correction methods, devices, equipment, storage media and program products

This application claims priority to the Chinese patent application with application number 202210660459.9 and the invention title "Color deviation correction method, device, equipment, storage medium and program product" submitted on June 13, 2022, the entire content of which is incorporated by reference. in this application.

Technical field

Embodiments of the present application relate to the field of image technology, and in particular to a color shift correction method, device, equipment, storage medium and program product.

Background technique

Currently, there are Standard Dynamic Range (SDR) displays and High Dynamic Range (HDR) displays to display pictures/videos. Among them, HDR displays can display more brightness information. Compared with The display effect is better on an SDR display. Therefore, the SDR image is usually converted into the corresponding HDR image and then displayed.

During the image conversion process, after the brightness is expanded, the hue and saturation of the image will change. For the skin color area of the image, it is easy to appear yellowish or reddish. Therefore, during the image conversion process, color cast needs to be performed. Correction.

Contents of the invention

The embodiments of the present application provide a color shift correction method, device, equipment, storage medium and program product. The technical solution is as follows:

On the one hand, embodiments of the present application provide a color shift correction method, which method includes:

Dynamically expand the brightness of the image parameters of the target SDR image to obtain the expanded image parameters, where the target SDR image contains portraits;

Based on the chromaticity information indicated by the image chromaticity parameter in the extended image parameters, select a local chromaticity parameter, where the local chromaticity parameter is a chromaticity parameter corresponding to the skin color area of the portrait;

Perform color deviation correction on the local chromaticity parameters to obtain corrected image parameters;

Based on the corrected image parameters, a target HDR image corresponding to the target SDR image is generated.

On the other hand, embodiments of the present application provide a color shift correction device, which includes:

A dynamic expansion module configured to dynamically expand the brightness of the image parameters of the target SDR image, where the target SDR image includes portraits, to obtain the expanded image parameters;

A parameter selection module configured to select a local chromaticity parameter based on the chromaticity information indicated by the image chromaticity parameter in the extended image parameter, where the local chromaticity parameter is a chromaticity parameter corresponding to the skin color area of the portrait;

A color deviation correction module configured to perform color deviation correction on the local chromaticity parameters to obtain corrected image parameters;

An image generation module configured to generate a target HDR image corresponding to the target SDR image based on the corrected image parameters.

On the other hand, embodiments of the present application provide a computer device. The computer device includes a processor and a memory. At least one instruction is stored in the memory. The at least one instruction is loaded and executed by the processor to implement Color shift correction method as described above.

On the other hand, embodiments of the present application provide a computer-readable storage medium in which at least one instruction is stored, and the at least one instruction is loaded and executed by a processor to implement the above aspects. color cast correction method.

On the other hand, embodiments of the present application provide a computer program product or computer program. The computer program product or computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the color shift correction method provided in various optional implementations of the above aspect.

Description of the drawings

Figure 1 shows a flow chart of a color shift correction method provided by an exemplary embodiment of the present application;

Figure 2 shows a flow chart of a color shift correction method provided by another exemplary embodiment of the present application;

Figure 3 shows a schematic diagram of the process of converting a target SDR image into a target HDR image provided by an exemplary embodiment of the present application;

Figure 4 shows a schematic diagram of the target chromaticity range in the Lab color space provided by an exemplary embodiment of the present application;

Figure 5 shows a flow chart of a color shift correction method provided by another exemplary embodiment of the present application;

Figure 6 shows a schematic diagram of the color shift correction process provided by an exemplary embodiment of the present application;

Figure 7 shows a structural block diagram of a color shift correction device provided by an embodiment of the present application.

Detailed ways

In order to facilitate understanding of the solutions provided by the embodiments of the present application, the terms involved in the embodiments of the present application are explained below.

SDR: SDR describes the dynamic range of the image/video through the traditional gamma gamma curve, where the SDR content is represented by 8bit. The color gamuts generally used for SDR content include Rec.601, Rec.709, and sRGB. Different color gamuts correspond to different gamma curves. SDR content is generally displayed on a screen with a brightness of 100nit.

HDR: HDR uses the hybrid log-gamma method (Hybrid Log-Gamma, HLG) or the perceptual quantizer (Perceptual Quantizer, PQ) to describe the dynamic range of the image/video. Compared with the gamma curve, the dynamic range that can be described is larger. In the embodiment of this application, the dynamic range refers to the brightness range. That is, the brightness range that can be displayed by the HDR standard is larger than the brightness range that can be displayed by the SDR standard. HDR content can be represented in 10bit, 12bit and 14bit. Currently, 10bit representation is mainly used. HDR content is generally displayed on a screen with a brightness of 1000nit-10000nit.

SDR2HDR: SDR to HDR, which converts traditional SDR content into HDR content.

Hunt effect: refers to the obvious change in the color appearance of an object as the overall brightness changes, that is, the chromaticity changes with the change in brightness. Although the traditional HSV/HSL/Lab/YUV/ICbCr color model can separate the brightness information and chroma information (hue and saturation) in the color to a certain extent, this separation is not perfect, and changes in brightness will still cause Bring about a change in color. In addition, the human eye itself The perception of color and brightness are closely related and difficult to completely separate.

YUV color coding: A color coding method commonly used in image processing components. Among them, Y represents brightness, U and V represent chroma.

YCbCr color encoding: YCbCr is a format of YUV color encoding, and YUV is often used to represent YCbCr.

Color deviation correction: the process of adjusting the chroma. Due to the Hunter effect, after the SDR image is converted into an HDR image, as the overall brightness changes, the hue and saturation of the color subjectively perceived by the human eye will also change, and an obvious color shift can be perceived. Therefore, it is necessary to Perform color cast correction.

Lab color space: CIE LAB, also called CIE L*a*b*. Lab consists of a brightness channel and two color channels. In the Lab color space, each color is represented by three components: L, a, and b, where L represents brightness; a represents the component from green to red; b represents the component from blue to yellow.

CIE XYZ: A color space created by the International Commission on Illumination (CIE) in 1931. The CIE XYZ space contains all colors that can be seen by the human eye. Each color is represented by three components: X, Y, and Z.

When converting SDR video to HDR images, the brightness will be increased due to brightness dynamic range expansion. After the brightness is increased, the corresponding chroma will change and the image saturation will increase. For most image areas, increasing saturation can make the image more natural and have better visual effects. However, for skin color areas, it will cause the skin color area to appear reddish, making the image less realistic and inconsistent with the real scene.

Therefore, in the embodiment of the present application, a color deviation correction method is provided. After dynamically expanding the brightness of the target SDR image, the color deviation correction will be further performed on the skin color area of the portrait, so that the converted HDR image is closer to reality. Scene; and only correct the skin color area in the image, which can avoid the impact on other areas and improve the visual effect.

The method provided by the embodiments of the present application can be applied to the process of converting a captured SDR video into an HDR video to perform color deviation correction on the skin color area in the video frame. It can also be used for real-time conversion of live broadcast scenes. During the live broadcast process, the live broadcast image can be converted in real time. During the process of converting the SDR image into an HDR image, the color cast of the skin color area can be corrected to improve the visual effect of live broadcast viewing. . In addition, the methods provided by the embodiments of the present application can also be used in other scenarios of converting SDR images/videos into HDR images/videos, which are not limited by the embodiments of the present application.

The method provided by the embodiment of the present application is executed by a computer device, where the computer device may have a video conversion function. It can be a smart phone, a tablet computer, a smart TV, a digital player, a laptop computer or a desktop computer, etc., which are not limited in the embodiments of the present application.

Please refer to FIG. 1 , which shows a flow chart of a color shift correction method provided by an exemplary embodiment of the present application. In this embodiment, the method is used in computer equipment as an example for description. The method includes the following steps.

Step 101: Dynamically expand the brightness of the image parameters of the target SDR image to obtain the expanded image parameters. The target SDR image includes portraits.

The method provided by the embodiment of the present application is suitable for the process of converting an SDR image to an HDR image, where the process of converting an SDR image to an HDR image involves a brightness dynamic expansion process. Brightness dynamic expansion refers to converting the SDR image into The brightness of the image extends to the brightness dynamic range supported by the HDR standard, making the image closer to the scene perceived by the human eye. After the brightness dynamic range is expanded, the image brightness increases, and due to the Henry effect, the chroma will increase accordingly, that is, the saturation will increase. In most scenes, increasing saturation will make the image more natural. However, for the skin color area of the portrait, the increase in saturation will cause the skin color to appear reddish, so color cast correction is required.

When the brightness of the target SDR image containing portraits is dynamically expanded, the expanded image parameters can be further corrected for color deviation, so that the converted image is closer to the scene perceived by the human eye.

Step 102: Select local chromaticity parameters based on the chromaticity information indicated by the image chromaticity parameters in the extended image parameters. The local chromaticity parameters are the chromaticity parameters corresponding to the skin color area of the portrait.

Normally, the skin color of the human body has a certain color range. In the embodiment of the present application, the computer device can identify the skin color area (including the face, neck, arms, etc.) in the target SDR image according to the corresponding color range of the skin color, so as to Color cast correction for skin tone areas.

In a possible implementation, the extended image parameters include image chromaticity parameters corresponding to each pixel point, which are used to indicate the chromaticity information of the pixel point. The computer device can determine whether the indicated pixel belongs to the color range corresponding to the skin color based on the chromaticity information of the indicated pixel. If the color range belonging to skin color is determined, it is determined that the pixel belongs to the skin color area. That is, the computer device can filter out the local chromaticity parameters according to the image chromaticity parameters and the color range of the skin color. The chromaticity information indicated by the local chromaticity parameters matches the color range of the skin color. The local chromaticity parameters correspond to the pixels composed of The area is the skin color area in the target SDR image, which can realize color deviation correction of the skin color area.

Compared with related technologies, color cast correction is performed on the entire image, and color cast correction can be performed only on the skin color area, which reduces the problem of unnatural skin color caused by increased saturation and avoids the impact on other areas.

Step 103: Perform color deviation correction on local chromaticity parameters to obtain corrected image parameters.

After selecting and obtaining the local chromaticity parameters, the computer device can perform color deviation correction on the local chromaticity parameters. In a possible implementation, since an increase in brightness will lead to an increase in saturation, the saturation can be reduced by reducing the chroma parameter.

In a possible implementation, the local chromaticity parameters can be reduced in the same proportion to complete the color deviation correction of the skin color area. In another possible implementation, since the chromaticity information indicated by different chromaticity parameters in the local chromaticity parameters is different, for example, some chromaticities are brighter and some chromaticities are duller, therefore, the computer device can also use the local chromaticity parameters according to the local chromaticity parameters. The chromaticity information indicated by the chromaticity parameters corresponding to different pixel points in the chromaticity parameters determines the reduction ratio, and the color deviation is corrected according to the corresponding reduction ratio. That is, different methods are used to correct color casts in different skin color areas to make the corrected image more natural and improve the visual effect.

Step 104: Generate a target HDR image corresponding to the target SDR image based on the corrected image parameters.

After obtaining the corrected image parameters, the computer device can further generate a target HDR image corresponding to the target SDR image based on the corrected image parameters to complete the conversion of the SDR image.

To sum up, in the embodiments of the present application, after dynamically expanding the brightness of the image parameters of the target SDR image, the computer device can determine the skin color in the target SDR image based on the chromaticity information indicated by the chromaticity parameters in the expanded image parameters. local chromaticity parameters corresponding to the area, thereby colorizing the local chromaticity parameters Bias correction to desaturate areas of skin tone. That is, the computer equipment can determine the skin color area based on the chroma information indicated by the image chroma parameters after brightness expansion, thereby correcting the color deviation of the skin color area, reducing the problem of high saturation in the skin color area caused by the increase in brightness, and improving the visual effect . And it can only correct the skin color area to avoid affecting other areas.

In the embodiment of the present application, before performing color deviation correction, the computer device first determines the skin color area in the image, thereby performing color deviation correction on the skin color area. In a possible implementation, the computer device determines whether it belongs to the skin color area based on the chromaticity information indicated by each pixel point. Exemplary embodiments will be described below.

Please refer to FIG. 2 , which shows a flow chart of a color shift correction method provided by another exemplary embodiment of the present application. In this embodiment, the method is used in computer equipment as an example for description. The method includes the following steps.

Step 201: Dynamically expand the brightness of the image parameters of the target SDR image to obtain the expanded image parameters.

Usually, the image parameters of the acquired target SDR image are nonlinear RGB parameters. In the process of dynamically expanding the brightness of the image parameters, as shown in Figure 3, the computer device first linearizes the nonlinear RGB parameters 301 to obtain the display linear RGB parameters. The display linear RGB parameters are used to indicate when displayed in the display screen. color and brightness information. After the linearization process, the color space conversion 302 will be performed to convert the display linear RGB parameters into the display linear XYZ parameters in the CIE Extended image parameters. The extended image parameters are the XYZ parameters after the brightness is dynamically expanded, which belongs to the CIE XYZ space.

Step 202: Perform color space conversion on the extended image parameters to obtain image chromaticity parameters in Lab space. The image chromaticity parameters are used to indicate image color information. The image chromaticity parameters include pixel chromaticity parameters corresponding to each pixel point.

In Lab space, the image parameters consist of the brightness parameter L and the chrominance parameters a, b. In the embodiment of the present application, based on the chromaticity information indicated by the chromaticity parameters a and b, it is determined whether it belongs to the skin color area.

In a possible implementation, the computer device first performs color space conversion on the extended image parameters, that is, converts the XYZ parameters corresponding to the extended image parameters in the CIE color space into L, a, and b parameters in the Lab color space.

Among them, the a and b parameters of each pixel constitute the image chromaticity parameters, which are used to indicate the image color information. That is, the image chromaticity parameters include the pixel chromaticity parameters corresponding to each pixel, respectively, including the a and b parameters corresponding to each pixel. .

Step 203: Based on the target chromaticity range, select the target chromaticity parameters from each pixel chromaticity parameter to obtain local chromaticity parameters, where the chromaticity information indicated by each target chromaticity parameter belongs to the target chromaticity range.

In a possible implementation, a corresponding target chromaticity range is preset for the skin color. Normally, the skin color area is located in the yellowish area, as shown in Figure 4, which is a schematic diagram of the Lab color space. According to the color attributes of the skin color, the target chromaticity range corresponding to the skin color area can be set in the Lab color space. For example, the target chromaticity range may be shown as area 401. When the pixel chromaticity parameter corresponding to the pixel points indicates When the color information is within the area corresponding to the target chromaticity range, it is determined to belong to the skin color area, and it can be determined as the target chromaticity parameter. Among them, based on the target chromaticity range, the method of selecting the target chromaticity parameters may include steps 203a-203b (not shown in the figure):

Step 203a, determine the angle parameter and length parameter corresponding to the chromaticity parameter of each pixel. The angle parameter and the length parameter are used to indicate the chromaticity position of the chromaticity in the Lab color space. The angle parameter and the length parameter are based on the a channel in the pixel chromaticity parameter. The parameters are determined with the b channel parameters.

In a possible implementation, when determining whether the pixel chromaticity parameter (a, b) belongs to the target chromaticity range, it can be determined whether the chromaticity position of the pixel chromaticity parameter (a, b) in the Lab color space is located The target chromaticity range is within the corresponding target area in the Lab color space.

Optionally, the computer device can determine the angle parameter and length parameter corresponding to the pixel chromaticity parameter based on the a and b channel parameters, thereby determining whether it is within the target chromaticity range based on the angle information and length information indicated by the angle parameter and the length parameter. within the corresponding target area.

Among them, the way to determine the angle parameter and length parameter according to the parameters a and b is as follows:

Among them, tanθ is the angle parameter, and r ² is the length parameter.

That is to say, in the embodiment of the present application, there is no need to directly calculate the angle information and length information, thereby reducing the complex calculation process and saving calculation overhead.

Step 203b: Based on the target angle range and the target length range, select the target chromaticity parameter to obtain the local chromaticity parameter. The angle parameter corresponding to the target chromaticity parameter belongs to the target angle range, and the length parameter belongs to the target length range.

In a possible implementation, after obtaining the angle parameter and length parameter corresponding to each pixel point, the judgment can be made based on the angle parameter and the length parameter. Optionally, the target area corresponding to the target chromaticity range is indicated by the target angle range and the target length range. When the angle parameter belongs to the target angle range and the length parameter belongs to the target length range, the computer device can determine that the corresponding pixel chromaticity parameter is the target chromaticity parameter, that is, it is determined to belong to the skin color area.

Illustratively, the target angle range can be 0<tanθ<2.14; the target length range can be 0<r ² <36. When a is 30 and b is 20 in the pixel chromaticity parameters corresponding to pixel point A, it can be determined that tanθ is 1.5, which belongs to the target angle range, and r ² is 13, which belongs to the target length range, and it is determined that pixel point A belongs to the skin color area. , and its corresponding pixel chromaticity parameter is the target chromaticity parameter.

After filtering based on the angle parameters and length parameters indicated by the pixel chromaticity parameters of each pixel, some target chromaticity parameters belonging to the target angle range and the target length range can be obtained to form local chromaticity parameters.

Step 204: Perform color deviation correction on local chromaticity parameters to obtain corrected image parameters.

After the local chromaticity parameters are selected and obtained, color deviation correction can be performed on each target chromaticity parameter in the local chromaticity parameters to obtain corrected corrected image parameters. For the color shift correction process, reference may be made to the following embodiments, which will not be described again in this embodiment.

It should be noted that after color shift correction, the computer equipment needs to convert the Lab parameters corresponding to each pixel into color space and re-convert them into parameters in the CIE XYZ color space. The corrected image parameters are the XYZ parameters after color space conversion.

Step 205: Perform color gamut conversion and non-linear conversion on the corrected image parameters to obtain a target HDR image under the HDR standard.

In the SDR2HDR process, as shown in Figure 3, after color shift correction 304, color gamut conversion 305 is performed to convert the XYZ parameters into HDR display linear RGB parameters under the Rec.2020 color space, where, the Rec.2020 color space Is the color space used in the HDR standard. After color gamut conversion, HDR display linear RGB parameters are obtained. The computer equipment also needs to perform HDR nonlinear conversion 306 to obtain nonlinear RGB parameters corresponding to the target HDR image and complete the SDR2HDR process.

In a possible implementation, when performing color deviation correction on the skin color area, the computer device can determine the color deviation correction method based on different chromaticity information. Exemplary embodiments will be described below.

Please refer to FIG. 5 , which shows a flow chart of a color shift correction method provided by another exemplary embodiment of the present application. In this embodiment, the method is used in computer equipment as an example for description. The method includes the following steps.

Step 501: Dynamically expand the brightness of the image parameters of the target SDR image to obtain the expanded image parameters.

Step 502: Select local chromaticity parameters based on the chromaticity information indicated by the image chromaticity parameters in the extended image parameters.

For the implementation of steps 501 to 502, reference can be made to the above steps 201 to 203, which will not be described again in this embodiment.

Step 503: Reduce each target chromaticity parameter in the local chromaticity parameters according to the target ratio to obtain corrected image parameters.

Since the increase in brightness causes an increase in saturation, the parameter value corresponding to the target chroma parameter can be reduced to achieve the purpose of reducing saturation. In a possible implementation, the computer device can reduce the parameter value corresponding to each target chromaticity parameter at a target ratio. As shown in the following formula:
a _new =a*S
b _new =b*S

Among them, S is the target ratio, and schematically, the target ratio can be 80%.

After color deviation correction is performed on the target chromaticity parameters, the updated parameters of parameter a and parameter b of the target chromaticity parameters are obtained, while other pixel chromaticity parameters that do not belong to the target chromaticity parameters remain unchanged, and the color deviation correction is completed. The corresponding image chromaticity parameters after correction are obtained, and the corrected image parameters are determined based on the corrected image chromaticity parameters. The corrected image parameters are the XYZ parameters after color space conversion of the corrected image chrominance parameters.

Step 504: Determine the correction weight based on the chromaticity position indicated by the target chromaticity parameter. The correction weight is positively correlated with the edge distance. The edge distance refers to the distance between the chromaticity position and the edge corresponding to the target chromaticity range.

For different target chromaticity parameters, the chromaticity positions indicated by the target chromaticity parameters in the Lab color space are different, and the color brightness corresponding to the target chromaticity parameters is different. If the same target ratio is used to compare each If the target chromaticity parameters are reduced, the image may be unnatural after color shift correction. Therefore, in another possible implementation, the computer device can determine the corresponding correction weight according to the chromaticity position indicated by the target chromaticity parameter, thereby determining the correction ratio based on the correction weight, and applying the correction weight to the target at different chromaticity positions. Colorimetric parameters are corrected at different scales. As shown in Figure 4, the chromaticity position indicated by the first target chromaticity parameter is the first position 402, and the chromaticity position indicated by the second target chromaticity parameter is the second position 403. The computer device is based on the target chromaticity. The chromaticity position indicated by the parameter determines the correction weight.

And when determining the correction weight, the chromaticity position is related to the distance between edges indicated by the target chromaticity range. If reduced at the same target ratio, the edges of the target chromaticity range will appear color-unsmooth. Schematically, as shown in Figure 4, the second position 403 indicated by the second target chromaticity parameter is located at the edge of the target area indicated by the target chromaticity range, which belongs to the target area. Therefore, the second target color will be Color deviation correction is performed on the chromaticity parameters. For example, the second target chromaticity parameter is reduced by 80%, and the pixel chromaticity parameters located outside the edge position and adjacent to the second target chromaticity parameter will not be adjusted. There is a certain gap between the two, and the color difference of the converted image is more obvious. Therefore, the computer equipment can determine the correction weight based on the distance between the chromaticity position and the edge indicated by the target chromaticity range. The closer to the edge, the lower the correction weight, and the color. The smaller the bias adjustment ratio. It should be noted that the computer device may determine the correction weight according to the minimum value of the distance between the chromaticity position and each edge indicated by the target chromaticity range.

In another possible implementation, the angle weight can be determined according to the angle parameter indicated by the target chromaticity parameter, and the length weight can be determined according to the length parameter indicated by the target chromaticity parameter, so that the correction weight can be determined based on the angle weight and the length weight. . This method may include steps 504a to 504b (not shown in the figure):

Step 504a: Determine the angle weight based on the angle parameter of the target chromaticity parameter. The angle weight is positively correlated with the angle difference. The angle difference refers to the difference between the angle parameter and the boundary parameter of the target angle range.

In a possible implementation, the angle weight is related to the angle difference, and the computer device can determine the corresponding relationship between the angle weight and the angle parameter based on the angle difference between the angle parameter and the boundary parameter of the target angle range, thereby avoiding angle edge color. Not smooth problem. The angle difference may be the minimum value of the difference between the angle parameter and the target angle range boundary parameter. That is, the corresponding relationship between the angle weight and the angle parameter is determined based on its distance from the target angle range. If the angle parameter differs greatly from the boundary parameter of the target angle range, it indicates that it is far away from the edge, so the adjustment can be made at a larger reduction ratio, that is, the angle weight is larger, and if the angle parameter differs from the boundary parameter of the target angle range, When it is smaller, it indicates that it is closer to the edge. In order to smooth the angular edge, it can be adjusted with a smaller reduction ratio, that is, the angular weight is smaller.

Optionally, when the angle parameter belongs to the first angle range, the angle weight corresponding to the angle parameter is determined based on the first angle curve, and the angle parameter under the first angle curve has a positive correlation with the angle weight.

When the angle parameter belongs to the first angle range, when the angle parameter gradually increases, the difference between it and the boundary parameter of the target angle range gradually increases. Therefore, the angle weight corresponding to the angle parameter can be determined by the first angle curve. An angle curve indicates the corresponding relationship between the angle parameter and the angle weight within the first angle range, which indicates that the angle parameter and the angle weight are positively correlated, and the angle weight is less than or equal to Weight threshold. When the angle parameter belongs to the first angle range, the larger the angle parameter is, the larger the edge difference from the target angle range is, and the greater the corresponding angle weight is.

Optionally, when the angle parameter belongs to the second angle range, the computer device determines the angle weight as the weight threshold, and the second angle range is greater than the first angle range.

When the angle parameter belongs to the second angle range, the second angle range is larger than the first angle range, and the difference between the angle parameter and the boundary parameter of the target angle range is relatively large. Therefore, it can be determined that the angle weight is the weight threshold, that is, it is determined The angle weight is the maximum value. Optional, the weight threshold can be 1.

Optionally, when the angle parameter belongs to the third angle range, the computer device determines the angle weight corresponding to the angle parameter based on the second angle curve. The angle parameter under the second angle curve has a negative correlation with the angle weight. The third angle The range is greater than the second angle range.

When the angle parameter belongs to the third angle range, the third angle range is larger than the second angle range. When the angle parameter gradually increases, the difference between the angle parameter and the boundary parameter of the target angle range gradually decreases. Therefore, The computer device may determine the angle weight corresponding to the angle parameter based on the second angle curve. The second angle curve indicates the corresponding relationship between the angle parameter and the angle weight within the third angle range, wherein the angle parameter and the angle weight are negatively correlated, and the angle The weight is less than or equal to the weight threshold. When the angle parameter belongs to the third angle range, the larger the angle parameter is, the smaller the edge difference with the target angle range is, and the smaller the corresponding angle weight is.

Schematically, the way to determine the angle weight based on the angle parameters is as follows:

Among them, f _θ is the angle weight. The first angle range is 0<tanθ<0.132; the second angle range is 0.132≤tanθ≤1.73; and the third angle range is 1.73<tanθ<2.14.

Step 504b: Determine the length weight based on the length parameter of the target chromaticity parameter. The length weight is positively correlated with the length difference. The length difference refers to the difference between the length parameter and the boundary parameter of the target length range.

Similarly, the length weight is related to the length difference. The computer device can determine the corresponding relationship between the length weight and the length parameter based on the length difference between the length parameter and the boundary parameter of the target length range, thereby avoiding the problem of uneven color at the length edge. The length difference may be the minimum value of the difference between the length parameter and the target length range boundary parameter. That is, the corresponding relationship between the length weight and the length parameter is determined based on its distance from the target length range. If the length parameter differs greatly from the boundary parameter of the target length range, it indicates that it is far away from the edge, so it can be adjusted with a larger reduction ratio, that is, the length weight is larger, and if the length parameter differs from the boundary parameter of the target length range, When it is smaller, it indicates that it is closer to the edge. In order to smooth the length edge, it can be adjusted by a smaller reduction ratio, that is, the length weight is smaller.

Optionally, when the length parameter belongs to the first length range, the computer device determines the length weight corresponding to the length parameter based on the first length curve, and the length parameter under the first length curve has a positive correlation with the length weight.

In the case where the length parameter belongs to the first length range, when the length parameter gradually increases, the difference between the length parameter and the boundary parameter of the target length range gradually increases. Therefore, the computer device can based on the first The length curve determines the length weight corresponding to the length parameter. The first length curve indicates the corresponding relationship between the length parameter and the length weight within the first length range, where the length parameter and the length weight are positively correlated, and the length weight is less than or equal to the weight threshold. . When the length parameter belongs to the first length range, the larger the length parameter is, the larger the edge difference from the target length range is, and the corresponding length weight is larger.

Optionally, when the length parameter belongs to the second length range, the length weight is determined to be the weight threshold, and the second length range is greater than the first length range.

When the length parameter belongs to the second length range, the second length range is greater than the first length range, and the difference between the length parameter and the boundary parameter of the target length range is relatively large. Therefore, the length weight can be determined to be the weight threshold, that is, it is determined The length weight is the maximum value. Optional, the weight threshold can be 1.

When the length parameter belongs to the third length range, the computer device determines the length weight corresponding to the length parameter based on the second length curve. The length parameter under the second length curve has a negative correlation with the length weight, and the third angle range is greater than the second length curve. angle range.

When the length parameter belongs to the third length range, the third length range is greater than the second length range. When the length parameter gradually increases, the difference between the length parameter and the boundary parameter of the target length range gradually decreases. Therefore, The length weight corresponding to the length parameter can be determined with a second length curve, and the second length curve indicates the corresponding relationship between the length parameter and the length weight within the third length range, wherein the length parameter and the length weight have a negative correlation, and the length weight is less than or equal to the weight threshold. When the length parameter belongs to the third length range, the larger the length parameter, the smaller the edge difference with the target length range, and the smaller the corresponding length weight.

Schematically, the way to determine the length weight based on the length parameter is as follows:

in, is the length weight. The first length range is 0<r ² <1, the second length range is 1≤r ² ≤30, and the third length range is 30<r ² <36.

Step 504c: Determine the correction weight based on the angle weight and the length weight.

After obtaining the angle weight and the correction weight, the computer device can determine the correction weight based on the angle weight and the length weight, and the correction weight can be the product of the angle weight and the length weight.

Step 505: Based on the correction weight, reduce the target chromaticity parameter to obtain the corrected image parameters. The correction weight has a positive correlation with the reduction ratio.

The computer device reduces the target chromaticity parameter according to the correction weight, where the correction weight is used to adjust the reduction ratio of the reduced parameter value. The higher the correction weight, the greater the reduction ratio.

Optional, based on the correction weight, the method of reducing the target chromaticity parameters is as follows:
S＝1-0.2*f _θ *f _r2
a _new =a*S
b _new =b*S

Among them, S is the reduction ratio.

Step 506: Perform color gamut conversion and non-linear conversion on the corrected image parameters to obtain a target HDR image under the HDR standard.

For the implementation of this step, reference can be made to the above-mentioned step 205, which will not be described again in this embodiment.

In this embodiment, the corresponding angle weight and length weight are determined respectively based on the angle parameter and the length parameter. When the angle parameter gradually approaches the target angle range, the angle weight becomes smaller, thereby smoothing the angle edge. When the length parameter gradually approaches the target length range, the angle weight becomes smaller. When , the length weight is smaller, thus smoothing the length edges, making the image after color shift adjustment more natural, and improving the visual effect.

In the above embodiment, when adjusting the target chromaticity parameter, the correction weight is determined only based on the chromaticity parameter, and the target chromaticity parameter is corrected based on the correction weight. In one possible case, when the pixel chromaticity parameter belongs to the target chromaticity range, the brightness value of the corresponding pixel may be very small, that is, the pixel is dark. In this case, it may not belong to the skin color area. Therefore, in a possible implementation, when correcting the target chromaticity parameter, a brightness parameter corresponding to the target chromaticity parameter is introduced, and the reduced target ratio is adjusted based on the brightness parameter. This method includes the following steps:

Step 1: Determine the brightness weight based on the target brightness parameter corresponding to each target chroma parameter in the local chroma parameter. The brightness weight has a positive correlation with the pixel brightness, and the brightness weight has a positive correlation with the reduction ratio.

The computer equipment can determine the brightness weight according to the target brightness parameter of the pixel corresponding to the target chromaticity parameter. When the brightness of the pixel is darker, the probability that it is a non-skin color area is greater. Therefore, the brightness weight is correspondingly lower, thus affecting the target. The chromaticity parameters are reduced by a lower percentage.

Optionally, the method of determining the brightness weight based on the target brightness parameter L is as follows:

Among them, f _L is the brightness weight, L _ref is the reference brightness value, and L _ref =100.

It should be noted that the brightness weight is less than or equal to 1.

Step 2: Based on the brightness weight and the correction weight, reduce each target chromaticity parameter in the local chromaticity parameters to obtain the corrected image parameters.

In a possible implementation, the computer device determines the reduction ratio corresponding to the target chromaticity parameter based on the brightness weight and the correction weight, thereby correcting the target chromaticity parameter, and determines the corrected image parameters based on the corrected parameters.

Based on the brightness weight and correction weight, the reduction ratio is determined as follows:
S＝1-0.2*f _θ *f _r2 *f _L

Schematically, as shown in Figure 6, the color shift correction process includes: first performing color space conversion 601, converting the XYZ parameters in the CIE XYZ space into Lab parameters in the Lab color space, and then performing color shift correction based on the Lab parameters. 602. In the color deviation correction process, first determine whether correction is required based on the chromaticity parameters a and b, that is, determine whether it belongs to the target angle range and the target length range based on the angle parameter and length parameter. If it belongs to the target angle range and If it falls within the target length range, it is determined that correction is required. During correction, the angle weight and length weight are determined respectively according to parameters a and b, and the brightness weight is determined according to parameter L. Finally, the reduction ratio is determined based on the angle weight, length weight, and brightness weight, and the color shift correction is performed with the reduction ratio. After correction, color space conversion 603 is performed to convert the Lab parameters in the Lab color space into XYZ parameters in the CIE XYZ space.

In this embodiment, when correcting the target chromaticity parameter, the brightness parameter is introduced at the same time, and the brightness weight is determined based on the brightness parameter, thereby further improving the accuracy of color deviation correction based on the brightness information.

Please refer to FIG. 7 , which shows a structural block diagram of a color shift correction device provided by an exemplary embodiment of the present application. The device includes:

The dynamic expansion module 701 is configured to dynamically expand the brightness of the image parameters of the target SDR image, where the target SDR image includes portraits, to obtain the expanded image parameters;

The parameter selection module 702 is configured to select a local chroma parameter based on the chroma information indicated by the image chroma parameter in the extended image parameter, where the local chroma parameter is a chroma parameter corresponding to the skin color area of the portrait. ;

The color deviation correction module 703 is configured to perform color deviation correction on the local chromaticity parameters to obtain corrected image parameters;

The image generation module 704 is configured to generate a target HDR image corresponding to the target SDR image based on the corrected image parameters.

Optionally, the parameter selection module 702 is also configured as:

Perform color space conversion on the extended image parameters to obtain image chromaticity parameters in Lab space. The image chromaticity parameters are used to indicate image color information. The image chromaticity parameters include pixel chromaticity parameters corresponding to each pixel point. ;

Based on the target chromaticity range, a target chromaticity parameter is selected from each of the pixel chromaticity parameters to obtain the local chromaticity parameter, wherein the chromaticity information indicated by each of the target chromaticity parameters belongs to the target chromaticity scope.

Optionally, the parameter selection module 702 is also configured as:

Determine the angle parameter and length parameter corresponding to each pixel chromaticity parameter. The angle parameter and the length parameter are used to indicate the chromaticity position of the chromaticity in the Lab color space. The angle parameter and the length The parameters are determined based on the a channel parameters and b channel parameters among the pixel chromaticity parameters;

Based on the target angle range and the target length range, the target chromaticity parameter is selected to obtain the local chromaticity parameter. The angle parameter corresponding to the target chromaticity parameter belongs to the target angle range, and the length parameter belongs to The target length range.

Optionally, the color shift correction module 703 is also configured to:

Reduce each of the target chromaticity parameters in the local chromaticity parameters by a target ratio to obtain the corrected image parameters;

or,

Based on the chromaticity position indicated by the target chromaticity parameter, a correction weight is determined. The correction weight is positively correlated with the edge distance. The edge distance refers to the correspondence between the chromaticity position and the target chromaticity range. The distance between the edges;

Based on the correction weight, the target chromaticity parameter is reduced to obtain the corrected image parameter, and the correction weight has a positive correlation with the reduction ratio.

Optionally, the color shift correction module 703 is also configured to:

Based on the angle parameter of the target chromaticity parameter, an angle weight is determined, the angle weight is the same as The angle difference is positively correlated, and the angle difference refers to the difference between the angle parameter and the boundary parameter of the target angle range;

Based on the length parameter of the target chromaticity parameter, a length weight is determined. The length weight is positively correlated with the length difference. The length difference refers to the distance between the length parameter and the boundary parameter of the target length range. difference;

The correction weight is determined based on the angle weight and the length weight.

Optionally, the color shift correction module 703 is also configured to:

When the angle parameter belongs to the first angle range, the angle weight corresponding to the angle parameter is determined based on the first angle curve, and the angle parameter under the first angle curve has a positive correlation with the angle weight;

In the case where the angle parameter belongs to a second angle range, determine the angle weight as a weight threshold, and the second angle range is greater than the first angle range;

When the angle parameter belongs to the third angle range, the angle weight corresponding to the angle parameter is determined based on the second angle curve, and the angle parameter under the second angle curve has a negative correlation with the angle weight, so The third angular range is larger than the second angular range.

Optionally, the color shift correction module 703 is also configured to:

When the length parameter belongs to the first length range, the length weight corresponding to the length parameter is determined based on the first length curve, and the length parameter under the first length curve is positively correlated with the length weight;

In the case where the length parameter belongs to a second length range, determine the length weight as a weight threshold, and the second length range is greater than the first length range;

When the length parameter belongs to the third length range, the length weight corresponding to the length parameter is determined based on the second length curve, and the length parameter under the second length curve has a negative correlation with the length weight, so The third angle range is larger than the second angle range.

Optionally, the color shift correction module 703 is also configured to:

Determine a brightness weight based on the target brightness parameter corresponding to each target chroma parameter in the local chroma parameters, the brightness weight is positively correlated with the pixel brightness, and the brightness weight is positively correlated with the reduction ratio;

Based on the brightness weight and the correction weight, each of the target chromaticity parameters in the local chromaticity parameters is reduced to obtain the corrected image parameters.

Optionally, the image generation module 704 is also configured to:

Perform color gamut conversion and non-linear conversion on the corrected image parameters to obtain the target HDR image under the HDR standard.

To sum up, in the embodiments of the present application, after dynamically expanding the brightness of the image parameters of the target SDR image, the computer device can determine the skin color in the target SDR image based on the chromaticity information indicated by the chromaticity parameters in the expanded image parameters. The local chroma parameters corresponding to the area are used to perform color deviation correction on the local chroma parameters to reduce the saturation of the skin color area. That is, the computer equipment can determine the skin color area based on the chroma information indicated by the image chroma parameters after brightness expansion, thereby correcting the color deviation of the skin color area, reducing the problem of high saturation in the skin color area caused by the increase in brightness, and improving the visual effect . and can only be used for skin color Correct the area to avoid affecting other areas.

It should be noted that the device provided in the above embodiments is only exemplified by the division of the above functional modules. In practical applications, the above function allocation can be completed by different functional modules as needed, that is, the internal structure of the device is divided into Different functional modules to complete all or part of the functions described above. In addition, the apparatus and method embodiments provided in the above embodiments belong to the same concept, and the implementation process can be found in the method embodiments, which will not be described again here.

An embodiment of the present application provides a computer device. The computer device includes a processor and a memory. At least one instruction is stored in the memory. The at least one instruction is loaded and executed by the processor to implement the above aspects. The color shift correction method described above.

Embodiments of the present application provide a computer-readable storage medium in which at least one instruction is stored, and the at least one instruction is loaded and executed by a processor to implement color shift correction as described above. method.

Embodiments of the present application provide a computer program product or computer program. The computer program product or computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the color shift correction method provided in various optional implementations of the above aspect.

Those skilled in the art should realize that in one or more of the above examples, the functions described in the embodiments of the present application can be implemented using hardware, software, firmware, or any combination thereof. When implemented using software, the functions may be stored on or transmitted over as one or more instructions or code on a computer-readable medium. Computer-readable media includes computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. Storage media can be any available media that can be accessed by a general purpose or special purpose computer.

Claims

A color shift correction method, wherein the method includes:

Dynamically expand the brightness of the image parameters of the target SDR image to obtain the expanded image parameters, where the target SDR image contains portraits;

Based on the chromaticity information indicated by the image chromaticity parameter in the extended image parameters, select a local chromaticity parameter, where the local chromaticity parameter is a chromaticity parameter corresponding to the skin color area of the portrait;

Perform color deviation correction on the local chromaticity parameters to obtain corrected image parameters;

Based on the corrected image parameters, a target HDR image corresponding to the target SDR image is generated.
The method according to claim 1, wherein the selecting local chromaticity parameters based on the chromaticity information indicated by the image chromaticity parameters in the extended image parameters includes:

Perform color space conversion on the extended image parameters to obtain image chromaticity parameters in Lab space. The image chromaticity parameters are used to indicate image color information. The image chromaticity parameters include pixel chromaticity parameters corresponding to each pixel point. ;

Based on the target chromaticity range, a target chromaticity parameter is selected from each of the pixel chromaticity parameters to obtain the local chromaticity parameter, wherein the chromaticity information indicated by each of the target chromaticity parameters belongs to the target chromaticity scope.
The method according to claim 2, wherein, based on the target chromaticity range, selecting a target chromaticity parameter from each of the pixel chromaticity parameters to obtain the local chromaticity parameters includes:

Determine the angle parameter and length parameter corresponding to each pixel chromaticity parameter. The angle parameter and the length parameter are used to indicate the chromaticity position of the chromaticity in the Lab color space. The angle parameter and the length The parameters are determined based on the a channel parameters and b channel parameters among the pixel chromaticity parameters;

Based on the target angle range and the target length range, the target chromaticity parameter is selected to obtain the local chromaticity parameter. The angle parameter corresponding to the target chromaticity parameter belongs to the target angle range, and the length parameter belongs to The target length range.
The method according to claim 3, wherein performing color deviation correction on the local chromaticity parameters to obtain corrected image parameters includes:

Reduce each of the target chromaticity parameters in the local chromaticity parameters by a target ratio to obtain the corrected image parameters;

or,

Based on the chromaticity position indicated by the target chromaticity parameter, a correction weight is determined. The correction weight is positively correlated with the edge distance. The edge distance refers to the correspondence between the chromaticity position and the target chromaticity range. The distance between the edges;

Based on the correction weight, the target chromaticity parameter is reduced to obtain the corrected image parameter, and the correction weight has a positive correlation with the reduction ratio.
The method of claim 4, wherein determining a correction weight based on the chromaticity position indicated by the target chromaticity parameter includes:

An angle weight is determined based on the angle parameter of the target chromaticity parameter. The angle weight is positively correlated with the angle difference. The angle difference refers to the difference between the angle parameter and the boundary parameter of the target angle range. difference;

Based on the length parameter of the target chromaticity parameter, a length weight is determined. The length weight is positively correlated with the length difference. The length difference refers to the distance between the length parameter and the boundary parameter of the target length range. difference;

The correction weight is determined based on the angle weight and the length weight.
The method of claim 5, wherein determining an angle weight based on the angle parameter of the target chromaticity parameter includes:

When the angle parameter belongs to the first angle range, the angle weight corresponding to the angle parameter is determined based on the first angle curve, and the angle parameter under the first angle curve has a positive correlation with the angle weight;

In the case where the angle parameter belongs to a second angle range, determine the angle weight as a weight threshold, and the second angle range is greater than the first angle range;

When the angle parameter belongs to the third angle range, the angle weight corresponding to the angle parameter is determined based on the second angle curve, and the angle parameter under the second angle curve has a negative correlation with the angle weight, so The third angle range is larger than the second angle range.
The method of claim 5, wherein determining a length weight based on the length parameter of the target chromaticity parameter includes:

When the length parameter belongs to the first length range, the length weight corresponding to the length parameter is determined based on the first length curve, and the length parameter under the first length curve is positively correlated with the length weight;

In the case where the length parameter belongs to a second length range, determine the length weight as a weight threshold, and the second length range is greater than the first length range;

When the length parameter belongs to the third length range, the length weight corresponding to the length parameter is determined based on the second length curve, and the length parameter under the second length curve has a negative correlation with the length weight, so The third angle range is larger than the second angle range.
The method according to claim 4, wherein performing color deviation correction on the local chromaticity parameters to obtain corrected image parameters includes:

Determine a brightness weight based on the target brightness parameter corresponding to each target chroma parameter in the local chroma parameters, the brightness weight is positively correlated with the pixel brightness, and the brightness weight is positively correlated with the reduction ratio;

Based on the brightness weight and the correction weight, each of the target chromaticity parameters in the local chromaticity parameters is reduced to obtain the corrected image parameters.
The method according to any one of claims 1 to 8, wherein generating a target HDR image corresponding to the target SDR image based on the corrected image parameters includes:

Perform color gamut conversion and non-linear conversion on the corrected image parameters to obtain the target HDR image under the HDR standard.
A color shift correction device, wherein the device includes:

A dynamic expansion module configured to dynamically expand the brightness of the image parameters of the target SDR image, where the target SDR image includes portraits, to obtain the expanded image parameters;

A parameter selection module configured to select a local chromaticity parameter based on the chromaticity information indicated by the image chromaticity parameter in the extended image parameter, where the local chromaticity parameter is a chromaticity parameter corresponding to the skin color area of the portrait;

A color deviation correction module configured to perform color deviation correction on the local chromaticity parameters to obtain corrected image parameters;

An image generation module configured to generate a target HDR image corresponding to the target SDR image based on the corrected image parameters.
A computer device, wherein the computer device includes a processor and a memory, at least one program is stored in the memory, and the at least one program is loaded and executed by the processor to implement any one of claims 1 to 9 Described color shift correction method.
A computer-readable storage medium, wherein at least one program is stored in the computer-readable storage medium, and the at least one program is loaded and executed by a processor to achieve the color shift as claimed in any one of claims 1 to 9. Correction method.
A computer program product, wherein the computer program product includes computer instructions stored in a computer-readable storage medium from which a processor of a computer device reads the computer instructions, The processor executes the computer instructions to implement the color shift correction method according to any one of claims 1 to 9.