Color model look-up table generation
The conversion of RAW data captured by a camera can have artifacts in smoothness of various hues for varying chroma. To optimize smoothness and color accuracy, transform coefficients defining conversion of a standard color model to a target color model are determined. The RAW data is converted to data in a standard color model and the data in the standard color model is converted to data in the target color model using the transform coefficients. The process is repeated for various lightness levels and combined into a look up table to efficiently convert RAW data to data in the target color model for various lightness levels.
1. A computer-implemented method comprising:
accessing a set of standard points in a standard color model;
determining a set of components of each standard point in the set of standard points, the set of components representing a color in the standard color model corresponding to the standard point;
receiving preference information for tuning the set of components;
applying the preference information to the set of components of each standard point to generate a target color model, the target color model including a different target point corresponding to each standard point;
determining transform coefficients for each pair of standard point and corresponding target point based on a relationship between the standard point and the corresponding target point by performing a polynomial fit operation to map the set of standard points to the target points, wherein the transform coefficients comprise polynomial coefficients of the polynomial fit operation;
accessing a set of RAW points in a RAW color model;
converting the set of RAW points to a second set of standard points in the standard color model; and
converting the second set of standard points to a set of optimized points in the target color model using the transform coefficients.
2. The computer-implemented method of claim 1 , wherein the standard color model is a standard RGB color model.
3. The computer-implemented method of claim 1 , wherein the set of components includes chroma and hue.
4. The computer-implemented method of claim 1 , wherein preference information is specific to chroma and hue.
5. The computer-implemented method of claim 1 , wherein preference information is specific to each point in the set of standard points.
6. The computer-implemented method of claim 1 , wherein the RAW color model is a RAW color model of a camera.
7. The computer-implemented method of claim 1 , wherein set of standard points and the second set of standard points do not overlap with at least one point.
8. The computer-implemented method of claim 7 , wherein the at least one point is converted to an optimized point in the target color model based on a plurality of polynomial fit operations performed on a plurality of nearest standard points.
9. The computer-implemented method of claim 8 , wherein the plurality of nearest standard points are points with the shortest distance to the at least one point.
10. A computer-implemented method comprising:
accessing a set of standard points in a standard color model;
determining a component of an initial point in the set of standard points;
receiving preference information for tuning the component;
applying the preference information to the component of the initial point to generate a target point in the target color model;
determining a transform coefficient for the target point and initial point based on a relationship between the initial point and the target point preference information by performing a polynomial fit operation to map the initial point to the target point, wherein the transform coefficient is based on one or more polynomial coefficients of the polynomial fit operation;
accessing a RAW point in a RAW color model;
converting the RAW point to a second point in the standard color model; and
converting the second point to an optimized point in the target color model using the transform coefficient.
11. The computer-implemented method of claim 10 , wherein the standard color model is a standard RGB color model.
12. The computer-implemented method of claim 10 , wherein the component is selected from a group consisting of: chroma and hue.
13. The computer-implemented method of claim 10 , wherein the set of standard points does not include the second point.
14. The computer-implemented method of claim 10 , wherein the second point is converted to the optimized point in the target color model based on a plurality of polynomial fit operations performed on a plurality of nearest standard points.
15. The computer-implemented method of claim 14 , wherein the plurality of nearest standard points are points with the shortest distance to the second point.
16. A camera, comprising:
a color model store that stores a set of standard points in a standard color model; and
a processor configured to:
access the set of standard points in the standard color model from the color model store,
determine a set of components of each standard point in the set of standard points, the set of components representing a color in the standard color model corresponding to the standard point,
receive preference information for tuning the set of components,
apply the preference information to the set of components of each standard point to generate a target color model, the target color model including a different point corresponding to each standard point,
determine transform coefficients for each pair of standard point and corresponding target point based on a relationship between the standard point and the corresponding target point by performing a polynomial fit operation to map the set of standard points to the target points, wherein the transform coefficients comprise polynomial coefficients of the polynomial fit operation,
access a set of RAW points in a RAW color model,
convert the set of RAW points to a second set of standard points in the standard color model, and
convert the second set of standard points to a set of optimized points in the target color model using the transform coefficients.