Integrated background and foreground tracking
AssigneeDisney Enterprises, Inc.
InventorYuecheng Zhang
Life of this patent
Estimating the term…
Abstract
Systems and methods for tracking the foreground and background objects in a video image sequence. The systems and methods providing for determining a camera model based on a first group of feature points extracted from a video frame, extracting three-dimensional (3D) information for the first group of feature points based on the camera model and a previous camera model for a previous video frame, reassigning feature points to the first group or a second group, based on mapping each feature point to a corresponding 3D model using the extracted 3D information for each feature point and determining a new camera model when a number of reassigned feature points is greater than a predetermined threshold.
Claims20
A method, comprising:
- determining a camera model based on a first group of feature points extracted from a video frame;
- extracting three-dimensional (3D) information for the first group of feature points and a second group of feature points based on the camera model and a previous camera model for a previous video frame;
- reassigning feature points to the first group or the second group, based on mapping each feature point to a corresponding 3D model using the extracted 3D information for each feature point;
- and
- determining a new camera model for the first group of feature points when a number of feature points reassigned is greater than a predetermined threshold.
A method, comprising:
- extracting a first group of feature points from a video frame;
- determining a camera model based on the first group of feature points;
- extracting a second group of feature points from a video frame;
- extracting three-dimensional (3D) information for the first group of feature points and the second group of feature points based on the camera model and a previous camera model for a previous video frame;
- reassigning feature points to the first group or the second group, based on mapping each feature point to a corresponding 3D model using the extracted 3D information for each feature point;
- and
- determining a new camera model for the first group of feature points when a number of feature points reassigned is greater than a predetermined threshold.
A video production system, comprising:
- a receiving component configured to receive a video frame;
- an extraction component configured to extract a first group of feature points and a second group of feature points from the video frame;
- a camera model estimator configured to determine a camera model based on the first group of feature points;
- a three-dimensional (3D) extraction component configured to extract 3D information for the first group of feature points and the second group of feature points based on the camera model and a previous camera model for a previous video frame;
- and
- an assignment component configured to reassign feature points to the first group or the second group, based on mapping each feature point to a corresponding 3D model using the extracted 3D information for each feature point, wherein the camera model estimator determines a new camera model for the first group of feature points when a number of feature points reassigned is greater than a predetermined threshold.
Dependencies are inferred from the claim text.