US-12621418-B2 - Color space mapping

US12621418B2US 12621418 B2US12621418 B2US 12621418B2US-12621418-B2

Abstract

Systems and methods for color space mapping are disclosed. First image data is received comprising first color information in a first color space, which can be a pixel-based color space. Second image data is generated based on the first image data, comprising converted second color information in a second color space based on the first color information. The second color space can be a pigment-based color space. The second image data is modified in the second color space using at least the converted second color information. Supplemental color data is generated based on the modified second image data. A modified image is generated in the first color space using the modified second image data and the generated supplemental color data, the modified image comprising modified color information in the first color space based at least in part on the supplemental color data. The second image data and the supplemental color data are generated using machine learning.

Inventors

Pablo Pernías Pascual de Pobil
Santiago Iglesias Navarro
Robert B. Moore
David N. Juboor

Assignees

DISNEY ENTERPRISES, INC.

Dates

Publication Date: 20260505
Application Date: 20240207

Claims (17)

1 . A computer-implemented method of color space mapping, the method comprising: receiving pixel-based image data for a first digital image comprising pixel color information in a pixel-based color space; generating, using the pixel-based image data, pigment-based image data, wherein the pigment-based image data comprises converted pigment color information in a pigment-based color space in a digital environment based on the pixel color information, wherein generating the pigment-based image data includes generating supplemental color data; modifying the pigment-based image data in the pigment-based color space using at least the converted pigment color information for the first digital image; and generating a modified pixel-based image using the modified pigment-based image data and the supplemental color data, wherein the modified pixel-based image is a second digital image and comprises modified color information in the pixel-based color space based at least in part on the supplemental color data.
2 . The computer-implemented method of claim 1 , wherein the pigment-based image data and the supplemental color data are generated using a machine learning model.
3 . The computer-implemented method of claim 1 , wherein the pixel-based color space uses a red-green-blue color model.
4 . The computer-implemented method of claim 1 , wherein the pigment-based color space uses a Kubelka-Munk color model.
5 . The computer-implemented method of claim 1 , wherein the supplemental color data comprises at least three channels of color data to be added to the converted pigment color information in an extended color space.
6 . The computer-implemented method of claim 1 , wherein generating the modified pixel-based image comprises determining pixel values by: converting a color value in the pigment-based color space to a corresponding pixel color value in the pixel-based color space; and modifying the corresponding pixel color value in the pixel-based color space based on the supplemental color data.
7 . A computer-implemented method of color space mapping, the method comprising: generating a training data set comprising multiple pixel-based color values and corresponding pigment-based color values; training, using the training data set, a machine learning model to generate supplemental color data and pigment-based color values based on pixel-based image data; receiving pixel-based image data comprising pixel color information in a pixel-based color space; generating, using the trained machine learning model, the supplemental color data and the pigment-based image data based on the received pixel-based image data, wherein the pigment-based image data comprises converted pigment color information in a pigment-based color space based on the pixel color information; modifying the pigment-based image data in the pigment-based color space using at least the converted pigment color information; and generating a modified pixel-based image using the modified pigment-based image data and the supplemental color data, wherein the modified pixel-based image comprises modified color information in the pixel-based color space based at least in part on the supplemental color data.
8 . A non-transitory computer-readable medium carrying instructions that, when executed by a processor, cause the processor to perform operations comprising: receive first image data comprising first color information in a first color space; generate, using the first image data, second image data, wherein the second image data comprises converted second color information in a second color space based on the first color information, wherein multiple color values in the second color space correlate to a same corresponding color value in the first color space, and wherein generating the second image data includes generating supplemental color data, wherein the second image data and the supplemental color data are generated using a machine learned model; modify the second image data in the second color space using at least the converted second color information; and generate a modified image in the first color space using the modified second image data and the supplemental color data, wherein the modified image in the first color space comprises modified color information in the first color space based at least in part on the supplemental color data.
9 . The non-transitory computer-readable medium of claim 8 , wherein the received first image data comprises a digital image.
10 . The non-transitory computer-readable medium of claim 8 , wherein the first color space is a pixel-based color space.
11 . The non-transitory computer-readable medium of claim 8 , wherein the second color space is a pigment-based color space.
12 . The non-transitory computer-readable medium of claim 8 , wherein the supplemental color data comprises at least three channels of color data to be added to the converted second color information in an extended color space.
13 . The non-transitory computer-readable medium of claim 8 , wherein the processor is further configured to perform operations comprising: generate a training data set comprising multiple color values in the first color space and corresponding color values in the second color space; and train, using the training data set, the machine learning model to generate supplemental color data in the second color space based on received image data in the first color space.
14 . The non-transitory computer-readable medium of claim 8 , wherein generating the modified image in the first color space comprises determining color values by: converting a color value in the second color space to a corresponding color value in the first color space; and modifying the corresponding color value in the first color space based on the supplemental color data.
15 . A computing system, comprising: at least one processor; and at least one non-transitory memory carrying instructions that, when executed by the at least one processor, cause the computing system to perform operations comprising: receive first image data comprising first color information in a first color space, wherein the first image data comprises a digital photograph; generate, using the first image data, second image data, wherein the second image data comprises converted second color information in a second color space based on the first color information, wherein multiple color values in the second color space correlate to a same corresponding color value in the first color space, and wherein generating the second image data includes generating supplemental color data; modify the second image data in the second color space using at least the converted second color information; and generate a modified image in the first color space using the modified second image data and the supplemental color data, wherein the modified image in the first color space comprises modified color information in the first color space based at least in part on the supplemental color data.
16 . The computing system of claim 15 , wherein the second image data and the supplemental color data are generated using a machine learning model.
17 . The computing system of claim 15 , wherein the first color space is a pixel-based color space.

Description

FIELD Described embodiments relate generally to color space mapping. BACKGROUND A digital image comprises visual elements (e.g., pixels) having various characteristics. Characteristics of visual elements can be represented numerically, such as color, intensity, or gray level. Digital images can be associated with various types, such as vector or raster type. Digital images can include photographs, such as unprocessed data captured using an image sensor of a digital camera. Digital images can be edited or manipulated in various ways. Color characteristics of digital image elements are represented using color values based on various color spaces. SUMMARY The following Summary is for illustrative purposes only and does not limit the scope of the technology disclosed in this document. In an example embodiment, a method of generating modified images in a first color space is disclosed. First image data is received comprising first color information in a first color space. The first color space can be a pixel-based color space, and the first image data can comprise a digital photograph. Second image data is generated using the first image data, the second image data including converted second color information in a second color space based on the first color information. The second color space can be a pigment-based color space. Multiple color values in the second color space correlate to a same corresponding color value in the first color space. Generating the second image data may include generating supplemental color data. Generating the second image data and/or the supplemental color data can be performed using a machine learning model. The supplemental color data can include at least three channels of color data to be added to the converted second color information in an extended color space. The second image data is modified in the second color space using the converted second color information and/or the supplemental color data. A modified image is generated in the first color space using the modified second image data and the generated supplemental color data, the modified image in the first color space including modified color information in the first color space based at least in part on the supplemental color data. In various embodiments, the method can include training the machine learning model. A training data set is generated comprising multiple color values in the first color space and corresponding color values in the second color space. The machine learning model is trained using the training dataset to generate supplemental color data and color values in the second color space based on received image data in the first color space. In various embodiments, generating the modified image in the first color space can include determining color values by converting a color value in the second color space to a corresponding color value in the first color space and modifying the corresponding color value in the first color space based on the supplemental color data. In an example embodiment, a non-transitory computer-readable medium is disclosed carrying instructions that, when executed by a processor and/or a computing system, cause performance of one or more methods disclosed herein. In an example embodiment, a computing system is disclosed comprising at least one processor and at least one non-transitory computer-readable medium carrying instructions configured to cause the computing system to perform one or more methods disclosed herein. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram illustrating a system flow for color space mapping. FIG. 2 is a block diagram illustrating a computing device for implementing a color space mapping system. FIG. 3 is a flow diagram illustrating a process performed using a color space mapping system. FIG. 4 is a flow diagram illustrating a process performed using a color space mapping system. DETAILED DESCRIPTION Digital images can use different color spaces to define color values (e.g., pixel values). As used herein, a “color space” can refer to a range of possible colors represented as corresponding sets of values (e.g., numerical values). Color spaces can be pixel-based, where pixels in a two-dimensional space defining a canvas are assigned a value corresponding to a color. An example of a pixel-based color space is a color space that uses RGB values, in which colors can be defined additively based on intensities of red, green, and blue elements. HSV values (hue, saturation, value) can also be used in a similar manner. Alternatively, color spaces can be pigment-based, where color values are assigned based on pigment-like behaviors or characteristics. A pigment-based color space can define color values in a digital image based on characteristics of pigments. For example, a pigment-based color space can allow for blending of colors in a digital image in a way that simulates the behavior of physical paint (e.g., watercolor, oil, or acrylic paint). Pigment-b