Share Email Print
cover

Proceedings Paper

Depth enhanced and content aware video stabilization
Author(s): A. Lindner; K. Atanassov; S. Goma
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

We propose a system that uses depth information for video stabilization. The system uses 2D-homographies as frame pair transforms that are estimated with keypoints at the depth of interest. This makes the estimation more robust as the points lie on a plane. The depth of interest can be determined automatically from the depth histogram, inferred from user input such as tap-to-focus, or selected by the user; i.e., tap-to-stabilize. The proposed system can stabilize videos on the fly in a single pass and is especially suited for mobile phones with multiple cameras that can compute depth maps automatically during image acquisition.

Paper Details

Date Published: 11 March 2015
PDF: 6 pages
Proc. SPIE 9411, Mobile Devices and Multimedia: Enabling Technologies, Algorithms, and Applications 2015, 941106 (11 March 2015); doi: 10.1117/12.2083443
Show Author Affiliations
A. Lindner, Qualcomm, Inc. (United States)
K. Atanassov, Qualcomm, Inc. (United States)
S. Goma, Qualcomm, Inc. (United States)


Published in SPIE Proceedings Vol. 9411:
Mobile Devices and Multimedia: Enabling Technologies, Algorithms, and Applications 2015
Reiner Creutzburg; David Akopian, Editor(s)

© SPIE. Terms of Use
Back to Top