Color conversion between color spaces using reduced dimension embeddings

Patent metadata

Field	Value
Publication number	US-12412314-B2
Application number	US-202418609524-A
Country	US
Kind code	B2
Filing date	Mar 19, 2024
Priority date	Aug 19, 2020
Publication date	Sep 9, 2025
Grant date	Sep 9, 2025

Abstract

Exemplary embodiments may provide an approach to converting multidimensional color data for an image encoded in a first color space into an intermediate form that is a single dimensional value. The exemplary embodiments may then decode the intermediate form value to produce an encoding of the color data that is encoded in a second color space that differs from the first color space. In this manner, the data for the image may be efficiently converted from an encoding in the first color space into an encoding in the second color space.

First claim

What is claimed is:

1. A method, comprising:
determining, by a neural network, a first color space of an image received as input, the image encoded in the first color space;
selecting, by the neural network based on the determined first color space of the image, a first encoder of a plurality of encoders to convert respective pixels of a first plurality of pixels of the image into a respective single-dimensional color value;
converting, by the first encoder, the respective pixels of the first plurality of pixels of the image into the respective single-dimensional color value;
selecting, by the neural network based on an output color space that is different than the first color space, a first decoder of a plurality of decoders to decode each respective single dimensional color value into a respective pixel of a second plurality of pixels in the output color space; and
decoding, by the first decoder, each respective single dimensional color value into the respective pixel of the second plurality of pixels in the output color space.

2. The method of claim 1 , wherein respective ones of the plurality of encoders and the plurality of decoders are trained using distinct color spaces.

3. The method of claim 1 , further comprising:
performing an image processing operation on the single-dimensional color values.

4. The method of claim 1 , wherein the second plurality of pixels in the output color space are compressed relative to the first plurality of pixels in the first color space.

5. The method of claim 1 , wherein a plurality of embedding values include the single-dimensional color values.

6. The method of claim 1 , wherein the image is a three-dimensional image, wherein a plurality of voxels of the three-dimensional image include the first plurality of pixels, wherein the first encoder encodes the plurality of voxels into the respective single-dimensional color values.

7. The method of claim 1 , wherein the first color space or the output color space is one of a RGB color space, an LAB color space, an HSV color space, a CMYK color space, a YUV color space, a HSL color space, an ICtCp color space, or a CIE color space.

8. A non-transitory computer-readable storage medium, the computer-readable storage medium including instructions that when executed by a processor, cause the processor to:
determine, by a neural network, a first color space of an image received as input, the image encoded in the first color space;
select, by the neural network based on the determined first color space of the image, a first encoder of a plurality of encoders to convert respective pixels of a first plurality of pixels of the image into a respective single-dimensional color value;
convert, by the first encoder, the respective pixels of the first plurality of pixels of the image into the respective single-dimensional color value;
select, by the neural network based on an output color space that is different than the first color space, a first decoder of a plurality of decoders to decode each respective single dimensional color value into a respective pixel of a second plurality of pixels in the output color space; and
decode, by the first decoder, each respective single dimensional color value into the respective pixel of the second plurality of pixels in the output color space.

9. The computer-readable storage medium of claim 8 , wherein the instructions further cause the processor to:
perform an image processing operation on the single-dimensional color values.

10. The computer-readable storage medium of claim 8 , wherein respective ones of the plurality of encoders and the plurality of decoders are trained using distinct color spaces.

11. The computer-readable storage medium of claim 8 , wherein the second plurality of pixels in the output color space are compressed relative to the first plurality of pixels in the first color space.

12. The computer-readable storage medium of claim 8 , wherein a plurality of embedding values include the single-dimensional color values.

13. The computer-readable storage medium of claim 8 , wherein the image is a three-dimensional image, wherein a plurality of voxels of the three-dimensional image include the first plurality of pixels, wherein the first encoder encodes the plurality of voxels into the respective single-dimensional color values.

14. The computer-readable storage medium of claim 8 , wherein the first color space or the output color space is one of a RGB color space, an LAB color space, an HSV color space, a CMYK color space, a YUV color space, a HSL color space, an ICtCp color space, or a CIE color space.

15. A computing apparatus comprising:
a processor; and
a memory storing instructions that, when executed by the processor, cause the processor to:
determine, by a neural network, a first color space of an image received as input, the image encoded in the first color space;
select, by the neural network based on the determined first color space of the image, a first encoder of a plurality of encoders to convert respective pixels of a first plurality of pixels of the image into a respective single-dimensional color value;
convert, by the first encoder, the respective pixels of the first plurality of pixels of the image into the respective single-dimensional color value;
select, by the neural network based on an output color space that is different than the first color space, a first decoder of a plurality of decoders to decode each respective single dimensional color value into a respective pixel of a second plurality of pixels in the output color space; and
decode, by the first decoder, each respective single dimensional color value into the respective pixel of the second plurality of pixels in the output color space.

16. The computing apparatus of claim 15 , wherein respective ones of the plurality of encoders and the plurality of decoders are trained using distinct color spaces.

17. The computing apparatus of claim 15 , wherein the instructions further cause the processor to:
perform an image processing operation on the single-dimensional color values.

18. The computing apparatus of claim 15 , wherein the second plurality of pixels in the output color space are compressed relative to the first plurality of pixels in the first color space.

19. The computing apparatus of claim 15 , wherein a plurality of embedding values include the single-dimensional color values.

20. The computing apparatus of claim 15 , wherein the first color space or the output color space is one of a RGB color space, an LAB color space, an HSV color space, a CMYK color space, a YUV color space, a HSL color space, an ICtCp color space, or a CIE color space.

Assignees

• Capital One Services Llc

Inventors

• Walters Austin Grant
• Goodsitt Jeremy Edward
• Pham Vincent

Classifications

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/0495

  Quantised networks; Sparse networks; Compressed networks · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/0455

  Auto-encoder networks; Encoder-decoder networks · CPC title

• G06F18/24

  Classification techniques · CPC title