Full-Frame Video Stabilization with Large Moving Object
Date Issued
2008
Date
2008
Author(s)
Huang, Wei-Ting
Abstract
This thesis presents an approach to post-processing casually captured videos to improvepparent camera movement. A lot of home videos have some problems about the artifactsike hand shaking when capturing without tripod. Video stabilization is an importantechnique to solve this problem. However, the technique does not work very well in someituations such as the larger foreground, incomplete background or other situations,etc. Inhis thesis,we propose a novel method of video stabilization to overcome the situations withhe larger foreground, or the feature points amount of the background is less then of foregroundnd some camera motion such as zoom in and zoom out.his system applies a method, optical flow, to estimate the motion vector of all pixelsetween each pair frames. Then, we use K-means clustering to group the similar motionectors of each frame. To select an adequate segment to estimate the global camera pathf video could obtain one more correct global camera path. After motion vector segmentationnd camera path estimation, we could stitch all of video frames to get a panoramand estimate the range of moving object we can recover by neighbor frames. Based on theackground panorama and moving object recovered range, we could find some new pathshich would lose information less. After the above operations, a full-stabilized video coulde achieved.
Subjects
home video
moving object detection
optical flow
polyline
video ROI
background segmentation
Type
thesis
File(s)![Thumbnail Image]()
Loading...
Name
ntu-97-R95922037-1.pdf
Size
23.32 KB
Format
Adobe PDF
Checksum
(MD5):115e20f989abd8955285280f0e4dd22c