Closed form method and system for large image matting
OwnerAlexandru Kuzmininventor
Life of this patent
Estimating the term…
Abstract
A method and system for large image matting, where alpha mask αR for a large image is built with the help of alpha mask αR found for the reduced image. In the current context, a large image is an image for which data structures required by the closed form method do not fit available memory. To overcome memory limitation the large image is divided into a plurality of smaller size sub-images. To find an alpha mask for every sub-image a sparse linear system is solved where the pixel values sampled from reduced image alpha mask αR are used as constraints on a sparse linear system.
Claims10
A method for a large image matting comprising:
- a) scaling down an original large image to receive a reduced image;
- b) building an alpha mask for said reduced image;
- c) upsampling said reduced image alpha mask to form an alpha grid;
- d) dividing said original large image and said alpha grid into a plurality of sub-images and alpha sub-grids;
- e) building an alpha mask for each said sub-image with the help of corresponding said alpha sub-grid by a modified closed form method where the nonempty pixel's values from corresponding said alpha sub-grid are used as constraints on a sparse linear system;
- f) building a final alpha mask for said original large image from said alpha masks built for sub-images.
A system a for large image matting comprising:
- a processing unit;
- a memory coupled with said processing unit storing instructions executable by processing unit;
- the instructions being operable to:
- a) scaling down an original large image to receive a reduced image;
- b) building an alpha mask for said reduced image;
- c) upsampling said reduced image alpha mask to form an alpha grid;
- d) dividing said original large image and said alpha grid into a plurality of sub-images and alpha sub-grids;
- e) building an alpha mask for each said sub-image with the help of corresponding said alpha sub-grid by a modified closed form method where the nonempty pixel's values from corresponding said alpha sub-grid are used as constraints on a sparse linear system;
- f) building a final alpha mask for said original large image from said alpha masks built for sub-images.
Dependencies are inferred from the claim text.