Method for video background subtraction using factorized matrix completion
View Patent ↗A method processes a video acquired of a scene by first aligning a group of video images using compressed domain motion information and then solving for a low rank component and a sparse component of the video. A homography map is computed from the motion information to determine image alignment parameters. The video images are then warped using the homography map to share a similar camera perspective. A Newton root step is followed to traverse separately Pareto curves of each low rank component and sparse component. The solving for the low rank component and the sparse component is repeated alternately until a termination condition is reached. Then, the low ranks component and the sparse component are outputted. The low rank component represents a background in the video, and the sparse component represents moving objects in the video.
1. A method for processing a video acquired of a scene, wherein the video includes a sequence of images, comprising the steps of:
solving for a low rank component of the video;
solving for a sparse component of the video;
following a Newton root finding step to traverse separately Pareto curves of each low rank component and sparse component to find successively better approximations to roots of a real-valued function;
repeating alternately the solving for the low rank component and the sparse component until a termination condition is reached; and
outputting the low rank component and the sparse component when the termination condition is reached, wherein the low rank component represents a background in the video, and the sparse component represents moving objects in the video, and wherein the steps are performed in a processor.
2. The method of claim 1 , wherein a restriction operator selects a subset of the sequence of images to be processed.
3. The method of claim 2 , wherein the low rank component is found by minimizing a proxy of a nuclear norm of a matrix based on a low rank factorization.
4. The method of claim 1 , wherein the video is acquired of the scene by a moving camera, and further comprising:
estimating the global motion parameters in the video to compensate for the camera motion.
5. The method of claim 4 , wherein the global motion parameters are estimated using encoded motion vectors from a video encoder or in a compressed video bitstream.
6. The method of claim 1 , wherein the processing is in real-time as the video is acquired.
7. The method of claim 1 , wherein the sparse component is determined using approximate message passing iterations with an adaptive threshold parameter.