Multi-frame simultaneous motion estimation and segmentation

This study proposes a new approach to jointly perform motion estimation and segmentation in the spatiotemporal domain. An energy functional that embeds the constraints of both the motion estimation and segmentation problems is minimized through the Euler-Lagrange equations yielding a set of descent equations executed via spatiotemporal level-sets representation. The resulting motion field is smooth yet boundary preserving and the motion segmentation map reflects the motion field enhancement.

[1]  Patrick Pérez,et al.  Joint Estimation-Segmentation of Optic Flow , 1998, ECCV.

[2]  Michael J. Black,et al.  Estimating Optical Flow in Segmented Images Using Variable-Order Parametric Models With Local Deformations , 1996, IEEE Trans. Pattern Anal. Mach. Intell..

[3]  Christoph Stiller,et al.  Object-based estimation of dense motion fields , 1997, IEEE Trans. Image Process..

[4]  Amar Mitiche,et al.  Motion tracking as spatio-temporal motion boundary detection , 2003, Robotics Auton. Syst..

[5]  Janusz Konrad,et al.  Video segmentation and occlusion detection over multiple frames , 2003, IS&T/SPIE Electronic Imaging.

[6]  Amar Mitiche,et al.  Tracking with simultaneous camera motion subtraction by level set spatio-temporal surface evolution , 2003, Proceedings 2003 International Conference on Image Processing (Cat. No.03CH37429).

[7]  A. Murat Tekalp,et al.  Simultaneous motion estimation and segmentation , 1997, IEEE Trans. Image Process..