Estimation of motion vector fields using tensor field filtering

Author

Westin, Carl-Fredrik ; Knutsson, Hans

Author_Institution

Dept. of Electr. Eng., Linkoping Univ., Sweden

Volume

2

fYear

1994

fDate

13-16 Nov 1994

Firstpage

237

Abstract

This paper presents a method for computation of two-dimensional motion vector fields from an image sequence. The magnitudes from a set of spatio-temporal quadrature filters are combined into a tensor description. The shape of the tensors describe locally the structure of the spatio-temporal neighbourhood and provides information about local velocity and if true flow or only normal flow is present. It is shown how normal flow estimates are combined into a true flow using linear filtering on this tensor field description

Keywords

estimation theory; filtering theory; image sequences; low-pass filters; motion estimation; quadrature mirror filters; tensors; image sequence; linear filtering; local velocity; low pass filtering; motion vector fields estimation; normal flow estimates; quadrature filter; spatio-temporal neighbourhood; spatio-temporal quadrature filters; tensor field description; tensor field filtering; tensors shape; true flow; two-dimensional motion vector fields; Computer vision; Equations; Filtering; Image motion analysis; Laboratories; Motion estimation; Nonlinear filters; Optical computing; Optical filters; Tensile stress;

fLanguage

English

Publisher

ieee

Conference_Titel

Image Processing, 1994. Proceedings. ICIP-94., IEEE International Conference

Conference_Location

Austin, TX

Print_ISBN

0-8186-6952-7

Type

conf

DOI

10.1109/ICIP.1994.413567

Filename

413567