Method of transmissivity-aware chroma keying

Patent metadata

Field	Value
Publication number	US-12388947-B2
Application number	US-202318155160-A
Country	US
Kind code	B2
Filing date	Jan 17, 2023
Priority date	Jan 18, 2022
Publication date	Aug 12, 2025
Grant date	Aug 12, 2025

Abstract

A method of transmissivity-aware chroma keying. The method includes: a) obtaining a first shot of at least one object in front of a first background or a first scene; b) obtaining a second shot of the at least one object in front of a second background or a second scene, which differs at least partially from the first background or the first scene; c) extracting the at least one object, using the first shot and the second shot.

First claim

What is claimed is:

1. A chroma keying method, the method comprising the following steps:
a) obtaining a first shot of an object in front of a first background, the first background being characterized by a first color;
b) obtaining a second shot of the object in front of a second background, the second background being characterized by a second color, the second color being different than the first color, so that the first and second backgrounds differ from each other with respect to color;
c) a modeler using varying qualities of appearance of the object in the first and second shots to estimate, for each of a plurality of pixels the modeler assigns to the object, a respective color and respective wavelength-dependent transmissivities to each of a respective color and transmissivity.

2. The method as recited in claim 1 , further comprising:
d) combining the plurality of pixels assigned to the object with a third background or a third scene, which differs at least partially from the first background and/or the second background.

3. The method as recited in claim 1 , wherein the method is used to generate mixed reality data for video blockages.

4. The method as recited in claim 1 , wherein the first shot and the second shot relate to the same image frame.

5. The method as recited in claim 1 , wherein the method is used for simulating at least one object, which blocks at least a portion of a view of a camera.

6. The method as recited in claim 1 , wherein the first and second background are the same as each other, except for their respective colors.

7. A non-transitory machine-readable storage medium on which is stored a computer program that is executable by a computer and that, when executed by the computer, causes the computer to perform a chroma keying method, the method comprising the following steps:
a) obtaining a first shot of an object in front of a first background, the first background being characterized by a first color;
b) obtaining a second shot of the object in front of a second background, the second background being characterized by a second color, the second color being different than the first color;
c) a modeler using varying qualities of appearance of the object in the first and second shots to estimate, for each of a plurality of pixels the modeler assigns to the object, a respective color and respective wavelength-dependent transmissivities to each of a respective color and transmissivity.

8. An object recognition system configured to perform the following for a chroma keying method:
a) obtain a first shot of an object in front of a first background, the first background being characterized by a first color;
b) obtain a second shot of the object in front of a second background, the second background being characterized by a second color, the second color being different than the first color;
c) implement a modeler that uses varying qualities of appearance of the object in the first and second shots to estimate, for each of a plurality of pixels the modeler assigns to the object, a respective color and respective wavelength-dependent transmissivities to each of a respective color and transmissivity.

Assignees

• Bosch Gmbh Robert

Inventors

• Klaehn Johann
• Lenor Stephan

Classifications

• H04N5/272

  Means for inserting a foreground image in a background image, i.e. inlay, outlay · CPC title

• H04N5/275

  Generation of keying signals · CPC title

• G06V10/82

  using neural networks · CPC title

• G06N3/08

  Learning methods · CPC title

• G06V10/42

  Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation · CPC title