Title :
Toward a robust parametric identification of motion on regions of arbitrary shape by nonlinear optimization
Author_Institution :
Centre Commun d´´Etudes de Telediffusion et Telecommun., Cesson-Sevigne Cedex, France
Abstract :
This paper highlights some problems related to the identification of parametric motion models in image sequences by the so-called differential methods. This often entails the bad behavior of the algorithms used, leading to poor results. Appropriate solutions are proposed. The differential algorithms are particularly effective for complex motion estimation, even of large magnitude
Keywords :
image segmentation; image sequences; motion estimation; optimisation; parameter estimation; algorithms; differential algorithms; differential methods; image region; image sequences; image shape; motion estimation; nonlinear optimization; parametric motion models; robust parametric identification; Finite element methods; Least squares methods; Motion estimation; Newton method; Parametric statistics; Polynomials; Recursive estimation; Robustness; Shape; Turning;
Conference_Titel :
Image Processing, 1995. Proceedings., International Conference on
Conference_Location :
Washington, DC
Print_ISBN :
0-8186-7310-9
DOI :
10.1109/ICIP.1995.529581