Variational motion segmentation based on discontinuity-preserving optical flow estimation

Bin Dai; Yuqiang Fang; Jinze Song

doi:10.1117/12.920952

13 January 2012 Variational motion segmentation based on discontinuity-preserving optical flow estimation

Bin Dai, Yuqiang Fang, Jinze Song

Proceedings Volume 8349, Fourth International Conference on Machine Vision (ICMV 2011): Machine Vision, Image Processing, and Pattern Analysis; 83490Q (2012) https://doi.org/10.1117/12.920952
Event: Fourth International Conference on Machine Vision (ICMV 11), 2011, Singapore, Singapore

Abstract

In this paper, a variational method for motion segmentation has been proposed, which is realized by first estimate discontinuity-preserving optical flow field and then segmenting. Specifically, the complementarity between estimation and segmentation was emphasized, so two main improvements were presented: firstly, a segmentation-oriented diffusion tensor has been suggested in the optical flow model, which combines the image and flow information and determine the boundaries more effectively; Secondly, according to the characteristic of vector field, a motion segmentation model with considerations of both motion boundaries and region information were established. The results were illustrated with examples of sequence images which show the usefulness of the proposed approach for various problems.

Citation Download Citation

Bin Dai, Yuqiang Fang, and Jinze Song "Variational motion segmentation based on discontinuity-preserving optical flow estimation", Proc. SPIE 8349, Fourth International Conference on Machine Vision (ICMV 2011): Machine Vision, Image Processing, and Pattern Analysis, 83490Q (13 January 2012); https://doi.org/10.1117/12.920952

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available